Conjugate molecule

Abstract

Methods for preparing and / or identifying multispecific antibody conjugates, such as bispecific antibody-drug conjugates, and the conjugated polyvalent molecules identified by these methods are provided herein. In particular, the present invention relates to a method for preparing a polyvalent conjugated polypeptide conjugated to a molecule of interest, comprising the step of combining a first polypeptide conjugated to the molecule of interest with at least one other polypeptide to form a polyvalent conjugated polypeptide conjugated to the molecule of interest. Intermediates, polypeptide constructs, kits, and uses are also provided.

Description

[Technical Field]

[0001] Field of Invention This invention relates to conjugate molecules such as antibody-drug conjugates. In particular, this invention relates to a method for identifying multi-specific antibody conjugates, such as bispecific antibody-drug conjugates and bispecific antibody-dye conjugates, and to the conjugate-multi-specific molecules identified by this method. Intermediates, kits, and uses are also provided.

[0002] Background of the Invention Antibody-drug conjugates (ADCs) are therapeutic agents containing an antibody (or antigen-binding fragment) and a drug. The drug is often referred to as the payload. The drug and antibody are typically linked by a linker. The linker is based on a chemical motif, usually including disulfide, hydrazone, or peptide (cleavable), or thioether (incleavable), and controls the distribution and delivery of the cytotoxic agent to target cells. Currently, more than 12 ADCs are approved as cancer treatments, and many more are in preclinical and clinical development.

[0003] More generally, drug conjugates can also be derived from non-antibody sources such as binding proteins or peptides, or antibody mimics such as cytokines or affibodies.

[0004] Another useful conjugate is the antibody dye conjugate, in which the antibody is labeled to facilitate detection in diagnostic settings or during preclinical drug discovery.

[0005] Bispecific antibodies are a common form of genetically modified antibody in which a single antibody-like molecule can bind to two different targets. For example, blinatumomab targets CD19 and CD3 and was approved by the US FDA in 2014 for the treatment of B-cell acute lymphoblastic leukemia. Another approved therapeutic bispecific antibody is emicizumab, which binds to coagulation factors IXa and X and is used to treat hemophilia A.

[0006] Bispecific antibodies, ADCs, and bispecific ADCs are reviewed in Shim Biomolecules. 2020 Mar; 10(3): 360.

[0007] Bispecific ADCs (BsADCs) are also discussed at https: / / www.biochempeg.com / article / 290.html (accessed June 6, 2023), where Figure 1 includes discussions on fragment expression, split intein trans-splicing, and subsequent drug conjugations for generating reactivated BsADCs.

[0008] Duckworth et al. (bioRxiv 2022.07.17.500350; doi: https: / / doi.org / 10.1101 / 2022.07.17.500350) describes the development of CD7 and CD33-targeted bispecific antibody-drug conjugates (BsADCs) for the treatment of acute myeloid leukemia by screening different bispecific ADC Fab assemblies targeting CD7 and CD33 against several cell lines. The bispecific constructs were created by ligating two Fab fragments via click chemistry and attaching a toxic payload to the ligated Fab fragments. The authors demonstrate that bispecific ADCs are cytotoxic to AML cells in vitro and are more selective than single-antigen-targeted ADCs in that they distinguish between tumor and healthy cells (either bone marrow or lymphoid).

[0009] Siegmund et al. (December 2016 Scientific Reports 6(1):39291) describe the development of a peptide-to-peptide ligation technique that enables the polymerization of tagged proteins catalyzed by SpyLigase using spontaneous isopeptide bond formation. The authors applied this technique to establish a modular antibody labeling method based on isopeptide bond formation between two recognition peptides, SpyTag and KTag. This labeling strategy allows the target reporting cargo to be attached to an antibody scaffold by chemically fusing the target reporting cargo with a KTag available via semi-automated solid-phase peptide synthesis (SPPS) and equipping the antibody with a SpyTag. The authors report successfully designing site-specific antibody-drug conjugates (ADCs) exhibiting sub-nanomolecal cytotoxicity using this strategy. The SpyTag-based system was also used by Kausar Alam et al. (Molecular Imaging and Biology volume 21, pages 54-66 (2019)). Their work describes site-specific fluorescent labeling of antibodies and diabodies using the SpyTag / SpyCatcher system for in vivo optical imaging.

[0010] International Publication No. 2022 / 200804 describes a polypeptide having a first binding domain at the N-terminus and a second binding domain at the C-terminus, wherein the first and second binding domains are separated by a structural domain, and each of the first and second binding domains is a catcher domain capable of forming an isopeptide linkage with a cognate peptide. The first catcher domain is linked to its cognate peptide tag by an isopeptide bond, and the second catcher domain is linked to its cognate peptide tag by an isopeptide bond. Each peptide tag is typically attached to an antigen-binding domain.

[0011] WO 2024 / 069180 pamphlet describes a multivalent protein scaffold that is useful as a therapeutic agent and useful for identifying new therapeutic compounds. The invention described therein also relates to a multi-domain polypeptide construct having a plurality of binding domains and structural domains. Also described are methods for identifying new candidate therapeutic agents using the provided multivalent protein scaffold, and the new therapeutic agents identified thereby. One described embodiment provides a polypeptide comprising a first binding domain at the N-terminus and a second binding domain at the C-terminus, wherein the first and second binding domains are separated by a structural domain. The first and second binding domains may be the same or different. In some embodiments, the first binding domain and the second binding domain are different antigen-binding domains. In this case, the construct is a bispecific construct. The structural domain may comprise or consist of the CutA1 protein. The CutA1 structural domain may be human CutA1. The CutA1 structural domain may comprise one or more substitutions or deletions compared to wild-type CutA1. In other embodiments, the structural domain comprises or consists of a cytokine derived from the TNF superfamily, such as, for example, TNF, OX40L, CD40L, or TL1A.

[0012] The great potential of bispecific antibodies and other multivalent and multispecific antibody formats has only just begun to be explored from a clinical perspective, but bispecific drug conjugates (e.g., ADCs) and multispecific drug conjugates (e.g., ADCs) are even newer research areas. There is a great need for improving the identification and development of multivalent, bispecific, and multispecific antibody molecules (and other targeting protein domains) conjugated to therapeutic or diagnostic molecules.

[0013] Summary of the Invention The present invention generally relates to the identification and development of conjugate molecules such as antibody conjugates and related antibody-based conjugates, in particular to multivalent antibody conjugates such as bispecific antibody-drug conjugates or bispecific antibody-dye conjugates. The present invention is partly based on the recognition that a combinatorial system that can be scaled up can be provided by a "conjugation first" approach in which the conjugation step is performed before the specificity is combined.

[0014] Existing techniques for generating conjugate molecules such as ADCs typically involve a first step of generating an antibody, followed by a second step of conjugating the target molecule (e.g., payload, drug, or dye) into the antibody. In contrast, the inventors have found that much larger scale and throughput can be achieved by first conjugating the target molecule into a portion of the antibody before forming a complete antibody. Furthermore, this method offers the storage and compatibility of pre-conjugated portions of constructs (e.g., antibodies), facilitating the formation of a complete construct, such as an antibody, through bulk pre-conjugation followed by numerous smaller reactions. This is particularly advantageous for screening and identifying useful multivalent conjugate molecules such as bispecific ADCs (BsADCs). The secondary nature of bispecificity material generation means that the methods described herein can provide improvements of several orders of magnitude, which can be further enhanced if triplicate antibodies or higher-order specificity are provided.

[0015] In a first aspect, the present invention provides a method for preparing a polyvalent polypeptide conjugated to a target molecule, comprising the step of combining a first polypeptide conjugated to a target molecule with at least one other polypeptide to form a polyvalent polypeptide conjugated to a target molecule. In some embodiments, the first polypeptide conjugated to a target molecule is combined with at least two other polypeptides, or at least three other polypeptides, or at least four other polypeptides to form a polyvalent polypeptide conjugated to a target molecule. The combination of the first polypeptide and one or more other polypeptides may be covalent or non-covalent.

[0016] A method for preparing a polyvalent bound polypeptide conjugated to a target molecule is, in some embodiments, (a) The step of conjugating the target molecule to the first polypeptide of a polyvalent conjugated polypeptide, and (b) The first polyvalent polypeptide conjugated from the polyvalent polypeptide obtained in step (a) is combined with at least one other polypeptide of the polyvalent polypeptide to form a polyvalent polypeptide conjugated to the molecule of the choice. It may include.

[0017] A method for preparing a polyvalent bound polypeptide conjugated to a target molecule is, in some embodiments, (a) The step of conjugating the target molecule to the first polypeptide of a polyvalent conjugated polypeptide, and (b) The first polyvalent polypeptide conjugated from the polyvalent polypeptide obtained in step (a) is combined with at least two other polypeptides of the polyvalent polypeptide to form a polyvalent polypeptide conjugated to the molecule of the choice. It may include.

[0018] The molecule of interest typically comprises or consists of a drug, label, or dye. In some embodiments, the molecule of interest is not a protein. In many embodiments of drug conjugates, the drug may be a non-protein drug, such as a small molecule drug, like a chemotherapeutic drug. Cytotoxic drugs, typically cytotoxic small molecule drugs, are typical drug components of ADCs in the art and are also typical molecules of interest in the present invention. Examples of cytotoxic drugs, as described elsewhere in this specification, include, but are not limited to, meltansine (also known as DM1), monomethyl auristatin F (MMAF), monomethyl auristatin E (MMAE), and deruxtecan. In certain embodiments, the small molecule drug is exatecan.

[0019] In some embodiments, the molecule of interest comprises or consists of nucleic acids, polynucleotides, or oligonucleotides. The nucleic acids, polynucleotides, or oligonucleotides may be DNA, RNA, XNA, LNA, or mixtures thereof. Typically, they are DNA or RNA. The nucleic acids, polynucleotides, or oligonucleotides may be single-stranded or double-stranded. In some embodiments, the oligonucleotides comprise or consist of 3 to 50 nucleotides (or nucleotide pairs in the case of double-stranded), for example, 5 to 30 nucleotides (or pairs). Antibody-oligonucleotide conjugates (AOCs) are often known in the art as a subset of ADCs.

[0020] In some embodiments, the molecule of interest is covalently conjugated to a first polypeptide, as is well known in the art.

[0021] In other embodiments, the molecule of interest is conjugated non-covalently to a first polypeptide, as is also known in the art. One non-limiting example of this method is a protamine (positively charged polypeptide) functionalized with a maleimide group covalently attached to a protein via a Cys thiol group. A nucleic acid (e.g., DNA) is then added to the protein-protamine conjugate, and the nucleic acid binds strongly to the protamine by electrostatic interaction. Thus, the nucleic acid is not directly attached to the protein, nor is it covalently attached to the protein. This basic method is also shown, for example, in Figure 1 of Dugal-Tessier et al., J Clin Med. 2021 Feb; 10(4): 838.

[0022] The molecule of interest can be conjugated to any portion of the first polypeptide. Typically, the conjugation occurs in step (a) of the present method. Certain benefits are derived when the molecule of interest is conjugated to a structural domain or to a linker region between a structural domain and a binding domain. In particular, conjugation to a structural domain or to a linker region allows for conjugation to a corresponding position on a drug candidate or drug molecule where the binding domain (e.g., SpyCatcher003) is replaced or removed by another binding domain (e.g., an antibody fragment), as described in International Publication 2022 / 200804. Thus, screening data on potential target combinations (e.g., identified using catcher-based format) can be used to provide information for designing drug candidates (e.g., identified conjugates [e.g., antigen-binding regions] directly connected to the structural domain), as shown in International Publication 2022 / 200804. In one embodiment, the conjugation of the molecule of interest to the CutA1 structural domain is provided.

[0023] A polyvalent conjugated polypeptide can bind to two or more epitopes. A polyvalent conjugated polypeptide typically comprises two or more target-binding regions, each specifically binding to an epitope, typically an antigen-binding domain, antigen-binding region, or antigen-binding construct. Each antigen-binding domain may be any antibody-based domain capable of immune-specific binding to an epitope. Each antigen-binding domain typically comprises six complementary delimiting regions (CDRs). Typically, each antigen-binding domain comprises six CDRs and a delimiting framework region, as known in the art. Typically, each antigen-binding domain comprises an immunoglobulin variable region containing CDRs and a framework region. Each antigen-binding domain may comprise two immunoglobulin variable regions, typically a heavy-chain variable region (VH) and a light-chain variable region (VL), as known in the art.

[0024] Polyvalent conjugated polypeptides typically include a binding domain (also called a binding region) based on the antibody's epitope-binding domain (i.e., variable domain) and include a CDR-delimited paratope. Many antibody-based domains, constructs, and forms are known in the art, including full-length antibodies, antigen-binding fragments of antibodies, and genetically modified antibody constructs. Antibody-based binding domains within the scope of the present invention include, but are not limited to, Fab fragments, F(ab')2, scFv fragments, scFv tandems, scFv-Fc, diabodies, scFv-CH3 (minibodies), scFab, human antibodies, and humanized antibodies. Polyvalent conjugated polypeptides may include, but are not limited to, aphibodies, nanobodies, or antibody-like scaffolds and non-scaffold conjugates containing DARPin; or, but are not limited to, cytokines containing TNF, TNFR, and receptor fragments or fusions such as etanercept, or target-binding fragments thereof.

[0025] A polyvalent polypeptide that binds to two epitopes is bivalent. A standard IgG antibody is bivalent. In some embodiments, a polyvalent binding polypeptide binds to three or more epitopes. A polyvalent polypeptide that binds to three epitopes is trivalent. In some embodiments, a polyvalent binding polypeptide binds to four or more epitopes. A polyvalent polypeptide that binds to four epitopes is tetravalent. In some embodiments, a polyvalent binding polypeptide binds to six epitopes. A polyvalent polypeptide that binds to six epitopes is hexavalent.

[0026] The epitopes to which a polyvalent binding polypeptide binds may be the same or different. A polyvalent binding polypeptide having multiple binding domains that bind to multiple copies of the same epitope is multivalent monospecific; for example, if it binds to two copies of the same epitope, it is divalent monospecific. A polyvalent binding polypeptide that binds to two different epitopes is divalent bispecific, but the double nomenclature is redundant, and such molecules are typically called bispecific. To avoid doubt, such a bispecific polypeptide targets each of two different epitopes at valence 1 ("1+1", such as a bispecific antibody derived from the hybridization of two IgG species), and is therefore divalent. Similarly, a bispecific polypeptide that targets each of two different epitopes at valence 2 ("2+2") is tetravalent. A bispecific polypeptide that targets each of two different epitopes at valence 3 ("3+3") is hexavalent. Similarly, polyvalent polypeptides can target each of their different target epitopes with different valencies or with the same valency. For example, a trivalent polypeptide may target the first epitope with valency 2 and the second epitope with valency 1 ("2+1"), or a trivalent polypeptide may target three different epitopes, each with valency 1 ("1+1+1"). Most natural antibodies are bivalent or polyvalent molecules containing identical antigen-binding sites, i.e., bivalent (or polyvalent) monospecific substances such as IgG, IgA, or IgM. An exception is some IgG4 molecules, which can exchange Fab arms due to the instability of their hinge region (half-antibody association).

[0027] A polyvalent binding polypeptide is multispecific if it binds to two or more different epitopes. In some embodiments, a polyvalent binding polypeptide is bispecific, i.e., it binds to two different epitopes. Bispecific antibodies are well known in the art. Bispecific antibodies, whose specificity is defined, are artificial molecules not found in nature. In some embodiments, a polyvalent binding polypeptide is triplicate or has higher-order specificity, having multiple binding regions that specifically bind to, for example, 4, 5, 6, or more different epitopes.

[0028] In some embodiments, the polyvalent conjugated polypeptide is a bispecific polypeptide. In some embodiments, the polyvalent conjugated polypeptide conjugated to the molecule of interest is a bispecific drug conjugate, e.g., a bispecific antibody-drug conjugate (bsADC). In some embodiments, the bispecific drug conjugate lacks one, some, or all of the domains of a conventional antibody (e.g., the dimerized IgG Fc region is replaced with a multimerizing domain, or the Fab region is replaced with a different binding domain, such as a nanobody, aphibody, DARPin, cytokine, naturally occurring ligand, or receptor fragment). In some embodiments, the bispecific drug conjugate contains only one or two domains of a conventional antibody (e.g., the dimerized IgG Fc region is replaced with a multimerizing domain, or the Fab region is replaced with a different binding domain, such as a nanobody, aphibody, DARPin, cytokine, naturally occurring ligand, or receptor fragment). In some embodiments, the drug component may not be a protein, but rather a cytotoxic drug (i.e., a "small molecule" drug) having a molecular weight of less than 1000 Da, for example.

[0029] If the polyvalent antigen-binding polypeptide is a bispecific polypeptide, such as a bispecific antibody, step (a) of the method may include conjugating the molecule of interest to the first polypeptide chain of the bispecific polypeptide. This may be, for example, one half of the bispecific antibody (e.g., one heavy chain and one light chain of a conventional full-length antibody). In this embodiment, step (b) may include combining the conjugated first polypeptide chain of the bispecific antibody with at least the second chain of the bispecific antibody (e.g., the second heavy chain and second light chain of a conventional full-length antibody), which may be the second half of the antibody, to form a bispecific antibody. Methods for conjugating a portion of an antibody in vitro are well known in the art.

[0030] In one embodiment of the preceding paragraph, a known technique called controlled Fab-arm exchange (cFAE), described in Labrijn et al. Nature Protocols volume 9, pages 2450-2463 (2014), is used to generate a bispecific antibody, typically bispecific IgG1, for example, a bispecific human IgG1 antibody. This method includes (i) separately expressing two parental IgG1 molecules containing a single matching point mutation in the CH3 domain; (ii) mixing the parental IgG1 molecules under redox conditions acceptable in vitro to allow recombination of the halves; (iii) removing a reducing agent to allow reoxidation of the interchain disulfide bond; and optionally (iv) analyzing the exchange efficiency and the final product, typically using a chromatography-based or mass spectrometry (MS)-based method. According to the present invention, one or both of the halves (obtained from step (i) and mixed in step (ii)) are the first polypeptide chain of a bispecific polypeptide. Accordingly, according to one embodiment of the present invention, an ADC is produced using the cFAE method, the cFAE method comprising the steps of (i) separately expressing two parental IgG1s containing a single matching point mutation in the CH3 domain, (ii) conjugating the molecule of interest onto at least one of the parental IgG1s, (iii) mixing the parental IgG1s (at least one containing the conjugate molecule of interest) under redox conditions acceptable in vitro to allow halve recombination, (iv) removing a reducing agent to allow reoxidation of the interchain disulfide bond, and optionally (v) analyzing the exchange efficiency and final product, typically using a chromatography-based or mass spectrometry (MS)-based method.

[0031] One application of the cFAE embodiment is shown in Barron et al., Int J Mol Sci. 2024 Feb; 25(4): 2097. In this embodiment, the method comprises the step of forming a bispecific ADC by applying cFAE to a first parental ADC conjugating to a first epitope (e.g., the antibody is conjugated with a cytotoxic small molecule such as MMAE) and a second parental unconjugated antibody conjugating to a second epitope. The cFAE results in an antibody comprising one conjugated half (derived from the conjugated parental antibody) and one unconjugated half (derived from the unconjugated parental antibody). Typically, this results in a biparatopic ADC, for example, shown in Figure 3 of Barron et al., which is specifically and expressly incorporated herein by reference. In certain embodiments, the first polypeptide of the polyvalent antigen-binding polypeptide comprises or consists of a polypeptide comprising a first binding domain, a second binding domain, and a structural domain. Each of these domains may contain more than one domain; for example, a structural domain may be an Fc region containing CH2 and CH3 domains. The first binding domain, the second binding domain, and the structural domain may also be referred to as the first binding region, the second binding region, and the structural region, respectively. Thus, in some embodiments, the first polypeptide of a multivalent antigen-binding polypeptide can be said to contain or consist of a polypeptide containing a first binding region, a second binding region, and a structural region. In such embodiments, the first polypeptide is a non-naturally occurring genetically engineered construct having a structural domain and a binding domain. These are typically expressed as recombinant multidomain polypeptides. The structural domain provides structurally defined support to the binding domain. Advantageously, in some embodiments, the structural domain can ensure that the binding domain has the desired orientation, typically so that both binding domains can bind to the target in cis-directing manner.Therefore, in some embodiments, the structure can present a single bonded surface.

[0032] An exemplary structural domain is the CutA1 protein, e.g., human CutA1. CutA1 possesses several beneficial features, including ease of expression, stability, oligomerization, and favorable orientation of the binding domains attached to the N-terminus and / or C-terminus. CutA1 as a structural domain is described in International Publication No. 2022 / 200804 and International Publication No. 2024 / 069180 (claiming priority to UK Patent Application No. 2214235.0).

[0033] Another exemplary structural domain is the Fc domain (also called the Fc region) of an antibody, such as the human Fc domain of human IgG. The human IgG domain may be an IgG1, IgG2, IgG3, or IgG4 Fc subtype. The Fc region includes CH2 and CH3 domains.

[0034] The CutA1 protein or Fc region may be a variant containing one or more modifications from the wild-type sequence, typically including substitutions, insertions, or deletions. Typically, the modifications are substitutions of 1 to 20 amino acid residues. Substitutions may involve replacing one or more cysteine ​​residues with non-cysteine ​​residues. Substitutions may involve replacing one or more non-cysteine ​​residues with cysteine ​​residues. The selection of cysteine ​​residues at a particular position may be useful for optimizing the conjugation of the target molecule. Substitutions may introduce non-natural amino acids. Non-natural amino acids may be useful for enabling the conjugation of the target molecule by biorthogonal chemistry, for example, by click chemistry. Non-natural amino acids (UAAs) are also known as non-proteinogenetic amino acids. Non-natural amino acids may be useful for enabling the conjugation of the target molecule by biorthogonal chemistry, for example, by click chemistry. Non-natural amino acids are known in the art and include D-amino acids, homoamino acids, N-methylamino acids, hydroxyproline (Hyp), beta-alanine, citrulline (Cit), ornithine (Orn), norleucine (Nle), 3-nitrotyrosine, nitroarginine, and pyroglutamic acid (Pyr). For example, propargyllysine is a non-natural amino acid that, when incorporated into proteins, can be used to attach commercially available fluorescent azide dyes via a copper-catalyzed alkyne-azide cycloaddition click reaction (also known as a click reaction).Other UAAs suitable for site-specific modification of polypeptide sequences include: 1: 3-(6-acetylnaphthalene-2-ylamino)-2-aminopropanoic acid (Anap), 2: (S)-1-carboxy-3-(7-hydroxy-2-oxo-2H-chromen-4-yl)propane-1-aminium (CouAA), 3: 3-(5-(dimethylamino)naphthalene-1-sulfonamide)propanoic acid (dansylalanine), 4: Nε -p-azidobenzyloxycarbonyllysine (PABK), 5:propargyl-L-lysine (PrK), 6:Nε-(1-methylcyclopropa-2-encarboxamide)lysine (CpK), 7:Nε-acryllysine (AcrK), 8:Nε-(cycloocta-2-in-1-yloxy)carbonyl)L-lysine (CoK), 9:bicyclo[6.1.0]nona-4-in-9-ylmethanollysine (BCNK), 1 0: trans-cycloocta-2-enlysine (2'-TCOK), 11: trans-cycloocta-4-enlysine (4'-TCOK), 12: dioxo-TCOlysine (DOTCOK), 13: 3-(2-cyclobuten-1-yl)propanoic acid (CbK), 14: Nε-5-norbornene-2-yloxycarbonyl-L-lysine (NBOK), 15: cyclooctinlysine (SCOK), 16: 5-norbornene Examples include n-2-oltyrosine (NOR), 17: cycloocta-2-inoltyrosine (COY), 18: (E)-2-(cycloocta-4-en-1-yloxyl)ethanoltyrosine (DS1 / 2), 19: azidohomoalanine (AHA), 20: homopropargylglycine (HPG), 21: azidonorleucine (ANL), and 22: Nε-2-azideoethyloxycarbonyl-L-lysine (NEAK).

[0035] In some embodiments, the first binding domain is located at the N-terminus of the polyvalent binding polypeptide, and the second binding domain is located at the C-terminus, with the first and second binding domains separated by a structural domain. In some embodiments, the first binding domain is connected to the second binding domain, and the second binding domain is connected to the structural domain. For example, a "catcher-catcher-structure" format, e.g., catcher-catcher-Fc format (e.g., shown as SpyCatcher003-DogCatcher-Fc in Example 8 and Figure 24). Typically, the second binding domain is connected to the N-terminus or C-terminus of the structural domain.

[0036] Accordingly, various embodiments of the present invention provide a binding (e.g., catcher) domain that is also separated by a structural domain (sometimes called a core protein) such as SpyCatcher003-Fc-DogCatcher and an alternative construct such as SpyCatcher003-DogCatcher-Fc (shown in Figures 1 and 24). Such platforms with alternative geometric shapes provide a useful method for evaluating the influence of conjugate configuration or valency on drug candidate behavior (as also shown in the comparison between different structural domains in Figures 11-14). Figure 24 shows that SpyCatcher003-DogCatcher-Fc is suitable for assembly with proteins comprising recombinantly introduced SpyTag / SpyTag003 or DogTag variants (hereinafter also referred to as "SpT protein" or "DgT protein," respectively), providing a novel protein platform useful for this screening method or related screening methods. This format is also suitable for pre-conjugation of small molecules (such as drug conjugation as shown below by fluorophore conjugation) followed by SpyCatcher / SpyTag-based and DogCatcher / DogTag-based protein assembly. Therefore, this format is a suitable polyvalent conjugation polypeptide in the context of the present invention. Figure 25 provides a drug screening based on a set of bispecific drug candidates assembled to SpyCatcher003-Fc-DogCatcher and SpyCatcher003-DogCatcher-Fc without payloads. This illustrates the difference in cytotoxicity due to changes in geometric shape.

[0037] Similarly, complexes for subsequent assembly can be created using knob-into-hole ("KiH") or related Fc heterodimerization techniques (for example, creating SpC-Fc / DgC-Fc heterodimers from fusion proteins of SpyCatcher or SpyCatcher003 [both referred to herein as "SpC"] or DogCatcher [also referred to herein as "DgC"] with an antibody Fc region, such as SpC-Fc and DgC-Fc). This specification describes how pre-Fc heterodimerization payload conjugations can be used to create pre-conjugated multivalent antigen-binding polypeptides with reduced DAR ("catcher-core" or "CC" constructs, i.e., constructs comprising at least one catcher domain and a core protein) (e.g., a pre-hybridization payload conjugation of a single SpC-Fc or DgC-Fc, resulting in an Fc hybrid from one half of the construct's conjugation) or two different linker-payload species (e.g., pre-hybridization conjugations of different linker-payload species with SpC-Fc and DgC-Fc). Similarly, post-Fc heterodimerization payload conjugations are also suitable for providing pre-conjugated complexes ready for assembly. When catcher technology is used as the binding domain, for example, if it is desired to maintain the same positional relationship between the conjugate and the protein tag relative to the Fc, then certain protein-protein assembly methods or related technologies for the sequential assembly of SpT conjugates may be preferred for knob-into-hole type Fc heterodimers such as SpC-Fc / SpC-TEV-Fc.

[0038] In some embodiments, KiH generation can be used as a step to introduce an additional screening dimension by conjugating, for example, Catcher-Fc and Catcher-TEV-Fc (or analogues, e.g., DogCatcher) with matching Fc mutations to different payloads, for example, to generate Catcher-Fc conjugated to MMAE and Catcher-TEV-Fc conjugated to exatecan, and then stepwise assembling them to produce bispecific, bipayload-type KiH. To create further diversity, multiple assembly techniques can be used in combination, such as catcher techniques (including catcher techniques that can be conditionally activated by protease digestion or light, or orthogonal catcher techniques such as SpyCatcher and DogCatcher), split-intein, sortase, or combinations of other techniques further described below.

[0039] The structural domain can also be modified to include useful mutations or modifications, such as “silencing” Fc mutations known in the art (e.g., “LALA”, “LAGA”, “LALAPG”), or to introduce or remove cysteine ​​residues or other residues for linker-payload attachment, thereby enabling the generation of catcher-core molecules with various DARs. For example, in in vitro assays, it may be desirable to evaluate silencing Fc-based assemblies instead of or compared to wild-type Fc-based assemblies in order to eliminate the potential influence of Fcγ receptors on the internal distribution of drug conjugates. We have demonstrated that additional cysteine, e.g., variations with eight Cys residues per dimer of the Fc-based catcher-core, to provide a greater number of small molecule conjugates facilitate drug-to-antibody ratios of up to eight. Such molecules are CC080 (SEQ ID NO: 53) and CC086 (SEQ ID NO: 54).

[0040] In general, the “conjugation-first” methods described herein can utilize various techniques sequentially or in addition to any single step to generate a wide variety of bispecificity, multispecificity, geometric shapes, forms, small molecule conjugations, combinations of small molecule conjugations, or other features.

[0041] In some embodiments, the first and second binding domains may be antigen-binding domains, for example, antibody-based domains containing six CDRs that form a paratope for binding to an epitope on a target.

[0042] Antigen-binding domains, regions, and constructs are well known in the art as described elsewhere herein, and include, for example, immunoglobulin-based forms such as Fab fragments, Fab' fragments, scFv, sdAb, or non-immunoglobulin-based antibody mimetic forms such as antikalin, fynomer, affimer, alphabody, DARPin, or avimer.

[0043] Linkers may be present between any two domains, or between all domains of a construct, or between any selected domains of a construct, such as linker peptides consisting of 2 to 30 amino acid residues, typically 5 to 25 amino acid residues.

[0044] In some embodiments, the first polypeptide of a polyvalent antigen-binding polypeptide comprises or consists of a polypeptide comprising a first binding domain, a second binding domain, and a structural domain. The first and second binding domains are inteins. Typically, the intein specificity of the first binding domain is different from that of the second binding domain. Inteins are self-processing domains found in organisms of all domains of life. These proteins carry out a process known as protein splicing, a multi-step biochemical reaction consisting of both the cleavage and formation of peptide bonds. While the endogenous substrate for protein splicing is a specific essential protein found in intein-containing host organisms, inteins can also function in exogenous contexts and can be used to manipulate virtually any polypeptide backbone.

[0045] In certain embodiments, the first polypeptide of the polyvalent antigen-binding polypeptide comprises or consists of a polypeptide comprising a first binding domain, a second binding domain, and a structural domain. The first and second binding domains are catcher domains, each capable of forming an isopeptide linkage with a congener peptide. Typically, the congener peptide for the first binding domain is different from the congener peptide for the second binding domain.

[0046] In some embodiments, the first and second binding domains are catcher domains, each capable of forming an isopeptide linkage with a congener peptide, and the congener peptide is the same for both the first and second binding domains. In embodiments where the congener peptide is the same for both the first and second binding domains, preferential, selective, or sequential binding or conjugation can be achieved by temporal or sequential control, spatial control, or spatiotemporal control.

[0047] If the first polypeptide contains two or more catcher domains and structural domains, at least one other polypeptide of the polyvalent antigen-binding polypeptide may contain an antigen-binding domain having a congener peptide capable of forming an isopeptide bond with at least one of the catcher domains. The isopeptide-forming peptide is typically linked to the antigen-binding domain by peptide linkage, and more typically, the peptide tag and antigen-binding domain are expressed as a single polypeptide chain.

[0048] Typically, the two catcher domains are distinct and form isopeptide bonds with different cognate peptides. Therefore, contact between a first polypeptide (containing two or more catcher domains) and an antigen-binding domain having a cognate peptide capable of forming an isopeptide bond with at least one of the catcher domains leads to the formation of the isopeptide bond, linking the antigen-binding domain to the structural domain via the catcher domains. In some embodiments, the at least one other polypeptide comprises two separate antigen-binding polypeptides, each having a cognate peptide capable of forming an isopeptide bond with one of the catcher domains. The two separate antigen-binding polypeptides can be combined with the first polypeptide simultaneously or sequentially.

[0049] In embodiments where both the first and second binding domains contain the same homologous peptide, preferential, selective, or sequential binding or conjugation can be achieved by temporal or sequential control, spatial control, or spatiotemporal control. This may be, for example, by activation or inactivation of one binding domain, and / or by competitive binding. Activation can be, for example, by enzymatic cleavage of inhibitory peptides (e.g., Driscoll et al, September 2023, bioRxiv 2023.08.31.555700; doi: https: / / doi.org / 10.1101 / 2023.08.31.555700, which is incorporated herein by reference in its entirety), or by photoactivation using a phototrigger for covalent bond formation in a photocage-binding domain created by site-specifically incorporating non-native residues such as coumarin-lysine into the reactive site (e.g., Rahikainen et al, 2023 "Visible light-induced specific protein reaction delineates early stages of cell adhesion", bioRxiv 2023.07.21.549850; doi: SpyCatcher003 is described at https: / / doi.org / 10.1101 / 2023.07.21.549850. This can be achieved by (the entire document is incorporated by reference). Further photoactivation methods using photocaged glutamate analogs are described in Yang et al, Angewendte Chemie Volume 62, Issue 40 October 2, 2023 (online published August 16, 2023) "Photoactivatable Protein with Genetically Encoded Photocaged Glutamic Acid". Alternatively, control of the availability of reactive sites can be achieved by removing any inhibitory domain, for example, by steric hindrance, binding competition, or blocking of reactive or catalytic residues.

[0050] In some embodiments, two antigen-binding domains, each containing the same isopeptide bond-forming tag, can both be precisely and selectively conjugated to a construct containing two of the same isopeptide bond-forming catcher domains (to which the tag binds), one of which is non-reactive until its reactivity is manifested using photoactivation. Typically, one of the catcher domains is non-reactive because it is caged at the reactive isopeptide bond-forming site by the presence of a non-native photoreactive residue, such as a coumarin-lysine (7-hydroxycoumarinlysine, "HCK") amino acid or a photocaged glutamate analog. Once the first tagged peptide has attached to the first catcher by forming an isopeptide bond between the first catcher and the tag, appropriate light (e.g., 405 nm light) is irradiated to release the photocage residue from the second catcher. The second catcher is then available for binding, and a second tagged peptide can be added. As used herein, “SpyTag” or “SpT” may refer to any suitable reactive form of SpyTag having beneficial reaction properties (e.g., SpyTag003), and in examples, generally refers to SpyTag003 (where “L2” uses the original SpyTag, and other SpT conjugates generally refer to SpyTag003). Similarly, “SpyCatcher” as used herein may refer to any suitable reactive form of SpyCatcher, and in examples, usually refers to SpyCatcher003 (see provided sequence numbers for details).

[0051] In some embodiments, two antigen-binding domains, each containing the same isopeptide bond-forming tag, can both be precisely and selectively conjugated to a construct of the present invention containing two identical isopeptide bond-forming catcher domains (to which the tag binds), one of which is non-reactive until reactivity is manifested using a site-specific protease. Typically, one of the catcher domains is non-reactive because it contains a non-reactive tag variant sequence fused to one of the catcher domains via a flexible linker containing a protease cleavage site. Once the first tagged peptide has attached to the first catcher by the formation of an isopeptide bond between the first catcher and the tag, a protease is added to release the non-reactive tag variant from the second catcher. The second catcher is then available for binding, and a second tagged peptide can be added.

[0052] In a particular embodiment, the non-reactive tag variant is SpyTag003 D117A (SpyTag003DA), fused to the C-terminus of a second SpyCatcher003 via a flexible linker containing a tobacco etch virus (TEV) protease cleavage site (Keeble et al., 2019 Proc. Natl. Acad. Sci. USA 116, 26523-26533; this document is incorporated herein by reference in its entirety). After cleavage at the TEV site, the SpyTag003DA peptide is free to dissociate, exposing the reactive lys of SpyCatcher003 and enabling reaction with the second supplied SpyTag linkage conjugate. Cleavage can be carried out using any suitable protease, such as super TEV protease.

[0053] In embodiments in which a first polypeptide containing two (or more) catcher domains and a target molecule conjugated is combined with two (or more) separate antigen-binding polypeptides, each having a congener peptide capable of forming an isopeptide bond with at least one of the catcher domains, the result is a molecule containing a structural domain, two catcher domains, and two antigen-binding domains, the two antigen-binding domains being connected to the catcher domains by isopeptide bonds. The two antigen-binding domains may be identical (forming a monospecific divalent molecule) or different (forming a bispecific molecule). If the combination of the two antigen-binding domains and the conjugated target molecule is determined to function well, for example, in one or more functional assays, the molecule can be further modified to remove the catcher domains, thereby providing an antigen-binding molecule essentially consisting of first and second antigen-binding domains separated by a structural domain. As described elsewhere herein, the structural domain is optionally CutA1, or an Fc domain, or a variant thereof. If the target molecule is conjugated to a structural domain or a linker connected to a structural domain, removing the catcher domain will retain the conjugated target molecule in the structural domain or linker, and typically, the position or orientation of the conjugated target molecule will not change. The combination of the binding domain and the conjugated target molecule can then be further developed, and it can be expected that the functions observed in the larger construct containing the catcher domain will be retained.

[0054] In embodiments in which a first polypeptide containing two catcher domains and a conjugated target molecule is combined with two separate antigen-binding polypeptides, each having a congener peptide capable of forming an isopeptide bond with at least one of the catcher domains, multiple conjugated first polypeptides of the polyvalent antigen-binding polypeptide can be generated in step (a) of the method. This generates multiple copies of the conjugated first polypeptide, each containing two or more catcher domains. Then, in step (b), these multiple copies of the multi-catcher polypeptide, each conjugated to a functional molecule of interest (e.g., a drug or dye), can be advantageously exposed to multiple different pairs of antigen-binding polypeptides, each having a congener peptide capable of forming an isopeptide bond with one of the catcher domains. As a result, a variety of different bispecific antigen-binding polypeptides are generated, each conjugated to a non-protein molecule of interest. This can be extended to a number of combinations of antigen-binding domains with a single target molecule, each of which can be evaluated for useful functions (e.g., cytotoxicity), and preferred compounds can be selected for further development.

[0055] Therefore, some embodiments are methods for preparing polyvalently bound polypeptides conjugated to a molecule of interest, (a) a step of conjugating a target molecule (e.g., a drug or dye or other payload) to a first polypeptide of a polyvalent binding polypeptide, wherein the first polypeptide comprises or consists of a polypeptide comprising a first binding domain, a second binding domain, and a structural domain, the first and second binding domains being catcher domains capable of forming isopeptide links with a congener peptide, and (b) A method comprising the step of combining the first polypeptide conjugated with the polyvalent polypeptide obtained in step (a) with at least two other polypeptides, each of which comprises or consists of an antigen-binding domain having a homologous peptide capable of forming an isopeptide bond with one of the catcher domains. The contact in step (b) is carried out under conditions suitable for isopeptide bond formation, thereby forming a polyvalent polypeptide conjugated to the molecule of interest.

[0056] In further embodiments, see Bhatta et al. 2021, mAbs, 13:1, 1859049 "Bispecific antibody target pair discovery by high-throughput phenotypic screening using in vitro combinatorial Fab libraries" https: / / www.tandfonline.com / doi / pdf / 10.1080 / 19420862.2020.1859049 (accessed June 16) th The non-covalent attachment of a protein to a tag, as described in 2023, can be used instead of one or more catchers and a congener peptide that forms an isopeptide bond. The non-covalent method may be suitable for simpler polyvalent polypeptides, such as divalent polypeptides that do not undergo polymerization.

[0057] The method of the present invention, in all embodiments, includes the step of conjugating a molecule of interest to a first polypeptide of a polyvalently linked polypeptide. This conjugated first polypeptide does not need to be used immediately and can be stored before being combined with one or more other polypeptides. In some embodiments, the conjugated first polypeptide is stored for at least one hour prior to step (b), and optionally the storage period is at least one day, at least one week, at least one month, or at least six months. Storage may be advantageous at refrigerated or frozen temperatures. Typically, storage is 10°C to -130°C. In some embodiments, storage is 10°C to -100°C. In some embodiments, the conjugated first polypeptide is stored at 8°C to -80°C, or 4°C to -20°C, or 4°C to 0°C.

[0058] A second aspect of the present invention provides a polyvalent conjugated polypeptide conjugated to a molecule of interest, which can be obtained or obtained by the method of the first aspect. In some embodiments, the polyvalent conjugated polypeptide is bispecific, triplicate, or has higher-order multiple specificity.

[0059] A third aspect of the present invention is a method for preparing a population of polyvalent antigen-binding proteins, wherein the members of the population have different antigen-binding domains, and each polyvalent antigen-binding protein is typically conjugated to a target molecule, which is a target non-protein molecule. (a) A step of preparing multiple polypeptides, each containing the target molecule conjugated thereto, Each polypeptide comprises a first binding domain, a second binding domain, and a structural domain, where the first and second binding domains are catcher domains capable of forming isopeptide links with congener peptides, and the congener peptides to the first binding domain are different from the congener peptides to the second binding domain, and (b) A method is provided comprising the step of contacting each of a plurality of polypeptides containing the target molecule conjugated thereto with a pair of antigen-binding polypeptides, each having a congener peptide capable of forming an isopeptide bond with one of the catcher domains, under conditions that enable the formation of an isopeptide bond between the catcher domain and the congener peptide, wherein different pairs of antigen-binding polypeptides are contacted with different first polypeptides containing the target small molecule conjugated thereto, resulting in a plurality of antigen-binding polypeptides having different antigen-binding characteristics and the same target small molecule conjugated thereto.

[0060] References to "2," "vs," and "first and second" can be increased accordingly for higher-order vaccines, with the presence of 3, 4, 5, or more binding domains.

[0061] In one embodiment of the third aspect of the method, the method is (c) A step of evaluating the characteristics of one or more antigen-binding polypeptides having different antigen-binding characteristics and small molecules of the same purpose conjugated thereto. It also includes.

[0062] In a variant of the third embodiment, the population of multivalent antigen-binding proteins includes molecules of different purposes conjugated to different members of the population. This allows for screening of multiple different payloads in addition to multiple combinations of binding domains. In some embodiments, there are two or more molecules of different purposes conjugated to different members of the population. In some embodiments, there are three or more molecules of different purposes conjugated to different members of the population. In some embodiments, there are four or more, five or more, ten or more, or twenty or more molecules of different purposes conjugated to different members of the population. In such embodiments, the method of the third embodiment may further include a step of evaluating one or more features of multiple antigen-binding polypeptides to which molecules of different purposes are attached and which have the same antigen-binding characteristics. Alternatively, in such embodiments, the method of the third embodiment may further include a step of evaluating one or more features of multiple antigen-binding polypeptides to which molecules of different purposes are attached and which have different antigen-binding characteristics, i.e., a comparison of multiple combinations of differences.

[0063] A fourth aspect of the present invention provides a library of multispecific antigen-binding polypeptides having different antigen-binding features and the same small molecule conjugated thereto, which can be obtained or obtained by the method of the third aspect.

[0064] A fifth aspect of the present invention provides a polyvalent conjugated polypeptide conjugated to a target molecule, or optionally to a target non-protein molecule, wherein the polyvalent conjugated polypeptide comprises a first binding domain, a second binding domain, and a structural domain, the first and second binding domains being catcher domains capable of forming isopeptide links with a congener peptide, and the congener peptide for the first binding domain is different from the congener peptide for the second binding domain.

[0065] In one embodiment of the fifth aspect, the first binding domain is located at the N-terminus, the second binding domain is located at the C-terminus, and the first and second binding domains are separated by a structural domain.

[0066] In another embodiment of the fifth aspect, the first binding domain is connected to a second binding domain, and the second binding domain is connected to a structural domain.

[0067] In a further embodiment of the fifth aspect, the catcher domains are covalently linked to their congener peptides by isopeptide linkage. In one embodiment, the congener peptides for the first catcher domain differ from those for the second catcher domain, as each congener peptide is covalently attached to a different antigen-binding domain. The covalent attachment between the peptide tag and the antigen-binding domain is typically a peptide linkage, and more typically, the peptide tag and antigen-binding domain are expressed as a single polypeptide chain.

[0068] In certain embodiments, a polyvalent conjugated polypeptide is provided that is suitable for conjugation with or conjugated to a molecule of interest. In some embodiments, a polyvalent conjugated polypeptide is provided that has at least 85%, at least 90%, at least 95%, at least 96%, at least 97%, at least 98%, or at least 99% identity with any exemplary polyvalent conjugated polypeptide described herein. In some embodiments, a polyvalent conjugated polypeptide is provided that has at least 85%, at least 90%, at least 95%, at least 96%, at least 97%, at least 98%, or at least 99% identity with SEQ ID NO: 47. In some embodiments, a polyvalent conjugated polypeptide is provided that has at least 85%, at least 90%, at least 95%, at least 96%, at least 97%, at least 98%, or at least 99% identity with SEQ ID NO: 52. In some embodiments, a polyvalent conjugated polypeptide having at least 85%, at least 90%, at least 95%, at least 96%, at least 97%, at least 98%, or at least 99% identity with SEQ ID NO: 53 is provided. In some embodiments, a polyvalent conjugated polypeptide having at least 85%, at least 90%, at least 95%, at least 96%, at least 97%, at least 98%, or at least 99% identity with SEQ ID NO: 54 is provided. In some embodiments, the identity percentage is at least 95%. In some embodiments, the identity percentage is at least 99%. Typically, the functionality of the exemplary sequences, particularly the functionality related to the methods described herein, is retained even when the sequence is altered.

[0069] A sixth aspect of the present invention provides a group of polyvalently bound polypeptides conjugated to a target molecule according to the second or fifth aspect.

[0070] A polyvalently bound polypeptide conjugated to the molecule of interest according to the second or fifth embodiment, or a group according to the sixth embodiment, can be formulated as a storage composition. Storage may be at a refrigerated or frozen temperature for at least one hour, at least one day, at least one week, at least one month, or at least six months. The formulation may contain one or more excipients. In some embodiments, one or more excipients may contain cryoprotectants.

[0071] A seventh aspect of the present invention provides a kit comprising a plurality of polyvalent conjugated polypeptides conjugated to a target molecule according to the second or fifth aspect, and a plurality of antigen-binding polypeptides, each having a homologous peptide capable of forming an isopeptide bond with one of the catcher domains. [Brief explanation of the drawing]

[0072] [Figure 1] This figure shows the generation of an ADC library. a) In current state-of-the-art technology, bispecific substances are generated by a scalable hybridization method and then conjugated with a payload. A payload conjugation step is performed for each generated ADC. b) In our method, the payload is first introduced into our 3+ partial assembly platform, and then bispecific diversity is generated through a scalable step. As a first step, a catcher-core library is conjugated to a toxic payload to generate a drug conjugate catcher-core library. Subsequently, the drug conjugate catcher-core library is assembled with tag A and tag B conjugate libraries to generate a drug conjugate / ADC panel. Compared to state-of-the-art methods, only one payload conjugation reaction is required to generate a diverse panel of different drug conjugates / ADCs. [Figure 2]This figure shows the shortening of the HsCutA1 terminus. a) Schematic diagrams of the coding sequences of different shortened HsCutA1 fused to SpyCatcher003 (SpC3) and DogCatcher (DgC) via exemplary linkers L1 and L2. The shortened terminus is denoted as 44-179 (SEQ ID NO: 16) and 60-171 (SEQ ID NO: 19), respectively. The sequences are compared in multiple sequence alignments between WT HsCutA1 (33-179) and HsCutA1 shortened to 44-179 and 60-171. b) Schematic diagram of the shortened HsCutA1 site. Residues 44-59 are shown at the N-terminus and residues 172-179 are shown at the C-terminus. Schematic diagram modeled from PDB ID: 2ZFH. c) Comparison of the expression of HsCutA1 44-179 and HsCutA1 60-171. We compared cell lysates of Escherichia coli (E. coli) to show how shortening altered protein expression. d) Apoptosis induced by SpC3-(GGGGS)3-[C75V,C96S]HsCutA144~179-(GGGGS)3-DgC (SEQ ID NO: 26) and SpC3-(GGGGS)3-[C75V,C96S]HsCutA160~171-(GGGGS)3-DgC (SEQ ID NO: 29), which form trivalent or hexavalent monospecific molecules variably assembled with L7 at either SpC3, DgC, or both. The percentage of viable Colo205 cells was determined 48 hours after initiation of treatment at the indicated dose. [Figure 3]This figure shows the removal of cysteine ​​from HsCutA1. a) Schematic representation of the naturally occurring cysteine ​​site in HsCutA1 that is to be removed. Schematic diagram modeled from PDB ID:2ZFH. b) Table of amino acid frequencies at sites similar to C75 and C96 of HsCutA1 after multiple sequence alignment (MSA) of 406 homologs. Amino acids with zero frequency are not shown. c) Comparison of the expression of HsCutA144-179 and HsCutA160-171 with natural WT cysteine ​​(C75, C96), alanine mutant cysteine ​​(C75A, C96A), and MSA-induced mutations (C75V, C96S). Cell lysates of Escherichia coli (E. coli) were compared to reveal all changes that mutagenesis brings to protein expression. d) Assemble with L7 of both SpC3 and DgC to form hexavalent monospecific molecules: SpC3-GGGGS-HsCutA144~179-GGGGS-DgC (SEQ ID NO: 15), SpC3-(GGGGS)3-[C75V,C96S]HsCutA144~179-(GGGGS)3-DgC (SEQ ID NO: 26), SpC3-(GGGGS)3-HsCutA1 Apoptosis induced by 60~171-(GGGGS)3-DgC (SEQ ID NO: 27), SpC3-(GGGGS)3--[C75A,C96A]HsCutA160~171-(GGGGS)3-DgC (SEQ ID NO: 28), and SpC3-(GGGGS)3--[C75V,C96S]HsCutA160~171-(GGGGS)3-DgC (SEQ ID NO: 29). The percentage of viable Colo205 cells was determined 48 hours after initiation of treatment with the indicated doses. [Figure 4]This figure shows a typical thiol-maleimide reaction with monomeric proteins, as well as various linker-payload structures. A) The thiol group of the monomeric protein reacts with the maleimide moiety of the paylaud-drinker compound to form a thiosuccinimodone paylaud-drinker adduct on the protein. B) Examples of linker-payload compounds conjugated to catcher-core proteins: 1-fluorescein-5-maleimide, 2-deruxtecan, 3-maleimidocaproyl-Val-Cit-PAB-DM1, 4-maleimidocaproyl-Val-Cit-PAB-MMAF. [Figure 5] This figure shows a schematic representation of the CutA1 structure (modeled from PDB ID: 2ZFH) and the position of the Cys mutation in a model of the fluorescein-5-thiosuccinimide-CutA1 conjugate. A) Spatial position of the Cys mutation in one monomer of the CutA1 structure. The Cys thiol is shown as a sphere. B) Model of the fluorescein-5-thiosuccinimide conjugate to the K82C mutant of CutA1 (SEQ ID NO: 31). All images were generated with PyMOL. [Figure 6] This figure shows fluorescein-5-maleimide conjugation for all Cys mutant CutA1 catcher-core variants (SEQ ID NOs: 40, 38, 41, 37, 39, 42) as well as the wild type (SEQ ID NO: 15; wild-type CutA1 is provided as SEQ ID NO: 1, and the truncated form incorporated into CC1 is provided as SEQ ID NO: 16) and the non-Cys (SEQ ID NO: 36) control. A) Dye / protein ratios for all mutants were calculated as described in the methods for samples after PD-10 purification. B) 12% SDS-PAGE gels of unconjugated and conjugated CutA1 mutants. Conjugation is little or no in the CC041 negative control, but all other mutants conjugate with fluorescein-5-maleimide. CC refers to the catcher-core protein gel band. [Figure 7]This figure shows the deconvolution ESI mass spectra of CC042 (SEQ ID NO: 38), as well as fluorescein-5-maleimide, deruxtecan, maleimidocaproyl-Val-Cit-PAB-DM1, and their maleimidocaproyl-Val-Cit-PAB-MMAF conjugates. A) Unconjugated catcher-core: The main peak is within 1 mass unit of the expected molecular weight. B) Catcher-core conjugated with fluorescein-5-maleimide: The main peak is approximately 18 Da larger than the expected molecular weight, which is likely due to hydrolysis of the thiosuccinimide ring after conjugation. C) Catcher-core conjugated with deruxtecan: The main peak is within 1 mass unit of the expected molecular weight. D) Catcher-core conjugated with maleimidocaproyl-Val-Cit-PAB-DM1: The main peak is within 1 mass unit of the expected molecular weight. E) Catcher-core conjugated with maleimidocaproyl-Val-Cit-PAB-MMAF: The main peak is within 1 mass unit of the expected molecular weight. [Figure 8] This figure shows SDS-PAGE (4-20% gradient) of fluorescein-5-maleimide conjugates and catcher / tag assemblies of CC042 (SEQ ID NO: 38) and CC041 non-Cys negative control (SEQ ID NO: 36). Only CC042 efficiently conjugates to fluorescein-5-maleimide. In CC041, a low level of background signal indicating conjugation is observed, suggesting that maleimide conjugation is primarily specific to Cys residues. Conjugated CC042 shows a clear change in migration within the gel when reacting with the DogTag conjugate (L7), the SpyTag conjugate (L8), or both conjugates simultaneously. Similar band profiles are observed in the CC041 control, suggesting that maleimide conjugation does not impair assembly efficiency. [Figure 9]This figure shows the catcher / tag assemblies of fluorescein-5-maleimide conjugation and CC068 (SEQ ID NO: 47) and CC060 (SEQ ID NO: 48) (Fc-catcher constructs) after one freeze / thaw cycle. In CC068, one of the three hinge region Cys residues is mutated to Ser(C230S), whereas in CC060, the hinge region is deleted, and therefore no surface-exposed Cys residues are present. Only CC068 conjugates to fluorescein-5-maleimide. In CC060, low levels of background conjugation are observed, suggesting that fluorescein-5-maleimide conjugation is primarily specific to the Cys residue. The fluorescein-5-maleimide conjugate CC068 exhibits clear and complete changes in its migration within the gel when reacted with a DogTag-tagged conjugate (L7), a SpyTag-tagged conjugate (L8), or both conjugates simultaneously. Similar band profiles were observed with CC060, suggesting that maleimide conjugation does not impair assembly efficiency. [Figure 10] This figure shows specific membrane binding and internal migration by RTK conjugates, single assemblies, and complete assemblies. HsCutA1 refers to CC7, and HsCutA1* refers to CC042 conjugated with fluorescein-5-maleimide. A) Staining with anti-His antibody shows that all compounds bound to the membrane and underwent internal migration of His-tagged conjugates L1 and L2, as well as the complete assembly with His-tagged CC7 (SEQ ID NO: 29). (Note that the ratio of His tags differs.) B) CutA1 staining of the complete assembly of CC7 with L1 and L2 conjugates confirms membrane binding and internal migration. C) Fluorescein-5-maleimide (F-5-M) conjugated CC042 (SEQ ID NO: 38) alone, as a single assembly and complete assembly, shows RTK-specific internal migration of all assemblies without off-target binding by HsCutA1*. [Figure 11]This figure shows cell viability data for HCT116 cells treated with protein assemblies of CC042 (SEQ ID NO: 38) and CC042 conjugated with DM1 (CC042DM1) having various ligands (L2, L4, L5, L6) at monomer concentrations of 50 nM (A), 10 nM (B), and 2 nM (C). Cell viability is presented as a heatmap. The left panel shows the difference in cell viability of the drug-conjugated CC042DM1 assembly compared to the corresponding CC042 assembly without drug conjugation. The center panel shows the cell viability of the CC042 assembly without drug conjugation, and the right panel shows the cell viability of the CC042DM1 assembly. All viability data were normalized to a sham control. [Figure 12] This figure shows cell viability data for HCT116 cells treated with protein assemblies of CC068 (SEQ ID NO: 47) and CC068 conjugated with deruxtecan (CC068D) at monomer concentrations of 20 nM (A), 4 nM (B), and 0.8 nM (C) with various ligands (L2, L4, L5, L6). Cell viability is presented as a heatmap. The left panel shows the difference in cell viability of the drug-conjugated CC068D assembly compared to the corresponding CC068 assembly without drug conjugation. The center panel shows the cell viability of the CC068 assembly without drug conjugation, and the right panel shows the cell viability of the CC068D assembly. All viability data were normalized to a sham control. [Figure 13]This figure shows cell viability data for HCT116 cells treated with protein assemblies of CC042 (SEQ ID NO: 38) and CC042 conjugated with MMAF (CC042MMAF) having various ligands (L2, L4, L5, L6) at monomer concentrations of 50 nM (A), 10 nM (B), and 2 nM (C) (C). Cell viability is presented as a heatmap. The left panel shows the difference in cell viability of the drug-conjugated CC042MMAF assembly compared to the corresponding CC042 assembly without drug conjugation. The center panel shows the cell viability of the CC042 assembly without drug conjugation, and the right panel shows the cell viability of the CC042MMAF assembly. All viability data were normalized to a sham control. [Figure 14] This figure shows cell viability data for HCT116 cells treated with protein assemblies of CC068 (SEQ ID NO: 47) and CC068 conjugated with MMAF (CC068MMAF) having various ligands (L2, L4, L5, L6) at monomer concentrations of 50 nM (A), 10 nM (B), and 2 nM (C). Cell viability is presented as a heatmap. The left panel shows the difference in cell viability of the drug-conjugated CC068MMAF assembly compared to the corresponding CC068 assembly without drug conjugation. The center panel shows the cell viability of the CC068 assembly without drug conjugation, and the right panel shows the cell viability of the CC068D assembly. All viability data were normalized to a sham control. [Figure 15]This is a schematic diagram of the coupling of SpyCatcher003 and DogCatcher with magnetic beads used in the assembly purification pipeline. Glutathione (GSH) conjugated magnetic beads are mixed with equimolar concentrations of GST-SpC3 and GST-DgC. The binding of the GST catchers to the beads is incubated at 25°C for 1 hour. The beads are then precipitated using a magnetic block and washed 3× with PBS or until a baseline absorbance reading at 280 nm is achieved. The GST-bound beads containing both catchers are retained for use during assembly purification. [Figure 16] This figure shows the SDS-PAGE (4-20% gradient) of glutathione conjugate bead preparations containing GST-SpC3 and GST-DgC. MagneGST resin was prepared using SDS loading buffer, and the presence of contaminating proteins in the stock was confirmed. W1-3: Washed with PBS. [Figure 17] This is a schematic diagram of another variation of catcher-based protein assembly and cleanup. SpC3-core-DgC is combined with SpT and DgT proteins with 1.4 × excess tagged protein relative to the catcher-core. The conjugation reaction is incubated at 25°C for 1 hour to allow complete capture of the tagged protein. Pre-prepared paramagnetic glutathione conjugate beads, conjugated to GST-SpC3 and GST-DgC as shown in Figure 15, are added so that each GST catcher is 2 × excess relative to the corresponding tagged protein. The sample is transferred to a magnetic block to capture the bead-GST-catcher-tag conjugate, and the supernatant containing the core-catcher-tag conjugate is retained for further analysis and downstream assays. [Figure 18]This figure shows SDS-PAGE (4-20% gradient) of protein assembly and MagneGST purification. The L7SpT:CC7:L2DgT protein assembly before purification (L7SpT:CC7:L2DgT(pre)) shows an excess of both conjugates, while the L7SpT:CC7:L2DgT protein assembly after purification (L7SpT:CC7:L2DgT(post)) shows only the fully assembled molecule. The GST-catcher + resin indicates that the excess conjugates were captured during MagneGST purification. (CC7 corresponds to Sequence ID No. 51). [Figure 19] This figure shows cytotoxicity data for a single cancer cell line treated with various conjugate assemblies assembled with unconjugated CC068 (SEQ ID NO: 47) at 0.5, 5, and 50 nM dimer concentrations. Each concentration was administered in triple duplicate doses. Conjugates were selected to cover a wide range of targets and were fabricated in various forms (including afibody, DARPin, scFv, and Fab). The units of the color scale represent cytotoxicity compared to the untreated negative control. Black indicates high cytotoxicity, and light gray indicates low cytotoxicity (see legend). The standard deviation is also shown in units of cytotoxicity. Black indicates a low standard deviation, and light gray indicates a higher standard deviation (see legend). X: Data excluded. Notations referring to conjugates and targets are shared within Figures 19–21 (though may differ from other figures). Conjugates with the same conjugate sequence but different tagging are assigned the same notation (e.g., Conjugate 1). [Figure 20]This figure shows cytotoxicity data for a single cancer cell line treated with various conjugate assemblies (specifically, MC-Val-Cit-PAB-MMAF) of CC068 (SEQ ID NO: 47) conjugated to MMAF at 0.5, 5, and 50 nM dimer concentrations. Each concentration was administered in triple overlapping doses. The conjugates were selected to cover a broad range of targets and were fabricated in various forms (including afibody, DARPin, scFv, and Fab). Payload-dependent cytotoxicity is evident when compared to Figure 19. The units of the color scale represent cytotoxicity compared to the untreated negative control. Black indicates high cytotoxicity, and light gray indicates low cytotoxicity (see legend). The standard deviation is also shown in units of cytotoxicity. Black indicates a low standard deviation, and light gray indicates a higher standard deviation (see legend). X: Data excluded. The notations for conjugates and targets are shared within Figures 19–21 (though they may differ in other figures). Conjugates with the same sequence but different tagging are assigned the same notation (e.g., Conjugate 1). [Figure 21]This figure shows cytotoxicity data for a single cancer cell line treated with various conjugate assemblies (specifically, Mal-PEG8-Val-Cit-PAB-MMAF) of CC068 (SEQ ID NO: 47) conjugated to PEG-MMAF at 0.5, 5, and 50 nM dimer concentrations. Each concentration was administered in triple duplicate doses. The conjugates were selected to cover a broad range of targets and were fabricated in various forms (including afibody, DARPin, scFv, and Fab). Compared with Figure 19, payload-dependent cytotoxicity is evident. Compared with Figure 20, this demonstrates the ability to rapidly test the effects of various linker-payload constructs. The units on the color scale are cytotoxicity compared to the untreated negative control. Black indicates high cytotoxicity, and light gray indicates low cytotoxicity (see legend). The standard deviation is also shown in units of cytotoxicity. Black indicates a low standard deviation, and light gray indicates a higher standard deviation (see legend). X: Data excluded. Notations referring to conjugates and targets are shared within Figures 19–21 (though may differ from other figures). Conjugates with the same sequence but different tagging are assigned the same notation (e.g., Conjugate 1). [Figure 22-1]Figure 22 consists of two parts, Figure 22A and Figure 22B (collectively referred to as Figure 22). Figure 22A demonstrates that the “conjugation-first” approach enables rapid, large-scale screening across a wide range of conjugates and / or target combinations. In a liquid handling setup, various conjugate combinations across a wide target range were assembled with CC068 (SEQ ID NO: 47) (specifically, Mal-PEG8-Val-Cit-PAB-MMAE) with or without payload conjugation with PEG-MMAE to generate a panel of approximately 800 drug candidates (of which approximately 400 were drug conjugates conjugated to PEG-MMAE). These were then administered in triple-duplicate, high-throughput drug in 384-well format to four different cell lines associated with a single indication, using 2.5 nM concentration assembly dimers. The total time from drug candidate generation to drug administration was <1 week. Experimental data were collected after 5 days using high-throughput fluorescence cell imaging. Left quadrant targets 2-5: Sections where SpT and DgT variants of the same conjugate are available. Includes tetravalent monospecific assemblies and "mirror image" assemblies (i.e., SpT conjugate A + DgT conjugate B compared to SpT conjugate B + DgT conjugate A). The units of the color scale are cytotoxicity compared to the untreated negative control. Black indicates high cytotoxicity, and light gray indicates low cytotoxicity (show legend). (n=1; show legend). X: Data excluded. Notations referring to conjugates and targets are shared within Figure 22A / B (but may differ from other figures). Conjugates with the same conjugate sequence but different tagging are assigned the same notation (e.g., conjugate 1). Figure 22B shows the standard deviation corresponding to Figure 22A. Standard deviation is shown in units of cytotoxicity. Black indicates a low standard deviation, and light gray indicates a higher standard deviation (show legend). X: Data excluded. The notations for conjugates and targets are shared with Figure 22A (though they may differ from those in other figures). Conjugates with the same sequence but different tagging are assigned the same notation (e.g., conjugate 1). [Figure 22-2] As stated above. [Figure 23]This figure shows that assembly clustering enables automated grouping of assemblies based on their activity across four cell lines. We collected values ​​for cytotoxicity, bispecific cytotoxicity difference (the difference in cytotoxicity between a bispecific assembly and the most cytotoxic corresponding single assembly containing one of the bispecific assembly conjugates), and ADC cytotoxicity difference (the difference in cytotoxicity between a PEG-MMAE conjugate assembly and an unconjugated assembly) for all PEG-MMAE conjugate assemblies across all four cell lines (12 features in total - rows in the clustered heatmap). Minmax normalization was applied to all 12 features to ensure values ​​were between 0 and 1 (see scale bars). Hierarchical clustering was then applied to the assemblies using the Euclidean distance metric and Ward's minimum variance method, progressively merging clusters. The maximum number of clusters was set to 30 (alternating dark gray, gray, and black bands in the "Cluster" row). Values ​​closer to 0 indicate higher cytotoxicity, and higher bispecificity and ADC cytotoxicity differences. The first three clusters (from left to right) contain assemblies that are active across three cell lines, while the last few clusters contain assemblies that show potent activity only in cell line 4. [Figure 24] This figure shows SDS-PAGEs of assemblies derived from various catcher-core proteins, with or without conjugation to the Alexa488 fluorophore molecule (specifically, Alexa488C5 maleimide). For CC076, CC080, and CC086, the gel on the left is Coomassie stained, and the gel on the right is imaged using the absorbance / fluorescence of the Alexa488 molecule. For Alexa488 conjugated samples, the complete single assembly and catcher-core protein annotations are shown. [Figure 25]This figure shows cytotoxicity data for a single cancer cell line treated with various conjugate assemblies of unconjugated CC068 (SEQ ID NO: 47) and CC076 (SEQ ID NO: 52) at dimer concentrations of 0.25, 2.5, and 25 nM. Comparing two assemblies with different orientations reveals differences in the cytotoxicity patterns for the same conjugate combination. This highlights the usefulness of screening with catcher-cores of different geometric shapes. The units on the color scale represent cytotoxicity compared to the untreated negative control. Black indicates higher cytotoxicity, and light gray indicates lower cytotoxicity (see legend). For standard deviation, black indicates a lower standard deviation, and light gray indicates a higher standard deviation (n=3; see legend). [Figure 26]This figure shows in vitro assays of exemplary bispecific antibodies bAb001 and bAb003 against target A+ / target B+ cell lines, either conjugated to or unconjugated to PEG-MMAE (specifically, Mal-PEG8-Val-Cit-PAB-MMAE) or PEG-exatecan (specifically, MC-PEG8-Val-Ala-PAB-exatecan). bAb001 is a construct of a binding domain directly fused to Fc with a geometric shape similar to CC068 (binding_domainA-Fc-binding_domainB), while bAb003 is a construct of a binding domain directly fused to Fc with a geometric shape similar to CC076 (binding_domainA-binding_domainB-Fc). A, B) Cell binding assays. Cancer cell lines were treated with protein at escalating concentrations up to a maximum antibody (dimer) concentration of 200 nM, allowing for binding saturation (n=2). Cell binding levels were quantified by incubation with anti-FC secondary antibodies conjugated to FITC, followed by quantification by flow cytometry. A) Percentage of sample bound to the cell surface, quantified by the percentage of cells in a positive gate set derived from negative controls. B) Cell binding quantified as mean fluorescence intensity (MFI), allowing for relative quantification of the amount of antibody bound to the cell surface. Nonlinear curve fitting was used to estimate the EC50 of the sample based on the MFI curve. The results are as follows: bAb001-no payload, 0.085 nM; bAb003-no payload, 0.14 nM; bAb001-PEG-MMAE, 0.10 nM; bAb001-PEG-exatecan, 0.077 nM. C) In vitro cytotoxicity assay. Cancer cell lines were treated with protein at escalating concentrations up to a maximum antibody concentration of 100 nM, and the cells were incubated for 5 days. Cell viability was quantified by Hoechst nucleus staining followed by readings from a Celigo Imaging Cytometer. The EC50 of the samples was estimated using nonlinear curve fitting.The results are as follows: bAb001 - No payload, 0.026 nM; bAb003 - No payload, 0.014 nM; bAb001 - PEG-MMAE, 0.00060 nM; bAb001 - PEG-Exatecan, Not applicable nM.

[0073] Detailed description of the invention The present invention generally relates to the identification and development of conjugates such as antibody conjugates, and more particularly to multivalent antibody conjugates such as multispecific antibody conjugates, for example, bispecific antibody-drug conjugates or bispecific antibody-dye conjugates.

[0074] The present invention is described with respect to specific embodiments and with reference to certain drawings, but the present invention is not limited thereto and is limited only by the claims. No reference numerals in the claims should be construed as limiting the scope. Of course, it should be understood that not all aspects or advantages are necessarily achieved according to any particular embodiment of the present invention. For example, a person skilled in the art will recognize that the present invention can be embodied or practiced in a manner that achieves or optimizes one advantage or group of advantages taught herein, but does not necessarily achieve other aspects or advantages that may be taught or suggested herein.

[0075] In addition, as used herein and in the appended claims, the singular forms “a,” “an,” and “the” refer to multiple objects unless the content clearly indicates otherwise. Thus, for example, a reference to “scaffold” includes two or more scaffolds, a reference to “oligomer” includes two or more such oligomers, and so on.

[0076] All publications, patents, and patent applications cited herein, whether above or below, are incorporated herein by reference in their entirety.

[0077] The present invention typically uses a polyvalent conjugated polypeptide conjugated to a target molecule, optionally a target non-protein molecule, wherein the polyvalent conjugated polypeptide comprises a first binding domain, a second binding domain, and a structural domain, the first and second binding domains being catcher domains capable of forming isopeptide links with a congener peptide, and the congener peptide for the first binding domain is different from the congener peptide for the second binding domain.

[0078] Structural Domain Structural domains provide structurally defined support for binding domains. In some embodiments, structural domains can ensure that binding domains have a desired orientation so that both binding domains can typically bind to a target in cis orientation. Thus, the construct can provide a single binding surface in some embodiments. In some embodiments, structural domains provide such structural support (e.g., preferred relative positioning of the N-terminus and C-terminus of a single monomer) by the tertiary structure of the monomer. In some embodiments, structural domains provide such structural support (e.g., preferred relative positioning of the N-terminus and / or C-terminus between monomers) by the tertiary structure of the monomer combined with the ancillary structure of the monomer.

[0079] The structural domain may be any polypeptide domain comprising a defined secondary structure, typically an alpha-helix or beta-sheet. In some particularly advantageous embodiments, the structural domain has, for example, substantially adjacent or adjacent N-terminuses and C-terminuses in the same spatial region. Attaching binding domains to the ends of the structural domain provides two substantially adjacent binding domains in a three-dimensional structure. In some embodiments, the N-terminuses and C-terminuses are oriented substantially in the same direction. Providing spatially adjacent N-terminuses and C-terminuses, as described elsewhere herein, results in binding domains located on the same plane of a construct or on the same plane of an oligomer containing multiple constructs. Thus, the construct typically provides a single binding surface. The construct typically provides a cis-oriented binding region.

[0080] A structural domain may contain a single polypeptide chain, or it may be composed of two or more separate polypeptide chains that together form a single structural domain complex. In some embodiments, two or more polypeptide chains with suitable features are identified and then typically fused by recombination to form a single polypeptide chain (i.e., a fusion protein), or chemically conjugated or bonded to form a single covalent molecule. Similarly, a structural domain may contain multiple domains, or may be referred to as a structural region, such as an Fc domain or Fc region (as further described below).

[0081] The structural domain is distinct from the two binding domains. Therefore, if the binding domain is a catcher polypeptide such as SpyCatcher, DogCatcher, or SnoopCatcher, the structural domain is not a catcher polypeptide.

[0082] A number of suitable structural domains are described in International Publication No. 2022 / 200804 and International Publication No. 2024 / 069180.

[0083] CutA1 CutA1, typically human CutA1, is a preferred structural domain. In some embodiments, the CutA1 protein is genetically engineered to contain one or more substitutions, insertions, or deletions compared to wild-type CutA1.

[0084] In a particular embodiment, the structural domain is human CutA1.

[0085] [ka]

[0086] In certain embodiments, the structural domain is a modified form of human CutA1 in which at least one cysteine ​​residue is replaced with a different amino acid residue. In certain embodiments, the structural domain is a modified and shortened form of human CutA1 in which at least one N-terminal and / or C-terminal amino acid residue is removed and at least one cysteine ​​residue is replaced with another amino acid residue.

[0087] In certain embodiments, CutA1, typically human CutA1, is genetically engineered to remove one or more cysteine ​​residues from its native sequence (as presented above). Typically, this involves substituting one or more cysteine ​​residues with one or more non-cysteine ​​residues. Removal of one or more cysteine ​​residues is beneficial because it allows for targeted cysteine ​​conjugation at non-native sites or in fusion proteins. While not theoretically bound, removing unpaired cysteines may be beneficial because they often hinder stability and downstream applications. In some embodiments, it is beneficial to remove one or more unpaired cysteines and reintroduce one or more cysteine ​​residues only at optimized target sites.

[0088] In some embodiments, one or more cysteine ​​residues of CutA1 are substituted with one or more alanine residues. In some embodiments, one or more cysteine ​​residues of CutA1 are substituted with one or more valine residues. In some embodiments, one or more cysteine ​​residues of CutA1 are substituted with one or more serine residues. In some embodiments, the substitution includes or consists of substituting two cysteines for two alanines, which is referred to herein as a "CACA" substitution. In some embodiments, the substitution includes or consists of substituting one cysteine ​​for valine and substituting one cysteine ​​for serine, which is referred to herein as a "CVCS" substitution. In some embodiments, the cysteine ​​residues at positions 75 and 96 of wild-type human CutA1 (e.g., SEQ ID NO: 1) are substituted with different residues. In some embodiments, human CutA1 is genetically engineered to have two cysteine ​​residues substituted, and the cysteine ​​substitutions include or consist of (i) C75A, C96A, or (ii) C75V, C96S. Thus, in some embodiments, the structural domain (or subunit monomer of the oligomeric core) of the polypeptide construct is a human CutA1 protein genetically engineered to have two cysteine ​​residues substituted, and the cysteine ​​substitutions include or consist of (i) C75A, C96A, or (ii) C75V, C96S. Figure 3 shows the generation and biological effects of polypeptide constructs containing such cysteine-substituted human CutA1 domains.

[0089] In some embodiments, it is beneficial to remove one or more unpaired cysteines and reintroduce one or more cysteine ​​residues only at optimized target positions. The reintroduced cysteines may be useful as conjugation sites for drugs or dyes, for example, when producing labeled antibodies (or antibody-type molecules) or antibody-drug conjugate ("ADC")-type molecules. Example 3 illustrates the substitution of non-cysteine ​​residues with cysteine ​​residues, and CutA1 with the newly introduced Cys residues has been observed to have higher conjugation efficiency when conjugating dyes or drugs to CutA1 compared to the wild type and negative control (CC041, i.e., CutA1 CACA genetically engineered to remove the natural cysteine). The newly introduced one or more cysteines can be substituted at preferred positions within the sequence. As illustrated in Example 13, exemplary positions in human CutA1 include one or more of V64, E78, K79, K82, E83, K91, Q102, K110, E114, F136, S139, F158, and Q166. In some embodiments, human CutA1 (e.g., those listed above) has cysteine ​​residue substitutions added to two, three, four, five, or six residues of E78, K82, Q102, E114, F136, and Q166.

[0090] In certain embodiments, CutA1, typically human CutA1, is genetically engineered to form a shortened CutA1 domain incorporated into the constructs of the present invention by deleting one or more residues from either or both ends of the native sequence. In some embodiments, one or more residues are deleted from the N-terminus of CutA1, typically human CutA1. Typically, 5 to 70 residues, e.g., 10 to 59 residues, are deleted from the N-terminus of CutA1, typically human CutA1. In some embodiments, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 50, 55, 59, 60, 61, 62, 63, 64, 65, or 66 residues are deleted from the N-terminus of CutA1. In some embodiments, the abbreviated CutA1 begins at residue 30 (i.e., 29 N-terminal residues are deleted). In some embodiments, the abbreviated CutA1 begins at residue 33 (i.e., 32 N-terminal residues are deleted). In some embodiments, the abbreviated CutA1 begins at residue 44 (i.e., 43 N-terminal residues are deleted). In some embodiments, the abbreviated CutA1 begins at residue 60 (i.e., 59 N-terminal residues are deleted). In some embodiments, the abbreviated CutA1 begins at residue 67 (i.e., 66 N-terminal residues are deleted), such as the "HsCutA1 67~171 CACA" direct fusion construct illustrated in Figure 23b of International Publication No. 2024 / 069180.

[0091] In some embodiments, one or more residues are deleted from the C-terminus of CutA1, typically human CutA1. The C-terminal deletion may be in place of or in addition to an N-terminal deletion. Typically, 5 to 20 residues are deleted from the C-terminus of CutA1, typically human CutA1, for example, 6 to 12 residues. In some embodiments, approximately 8 residues are deleted. In human CutA1, deleting 8 residues results in a shortened protein having a C-terminus where residue 171 (valine) of the wild-type sequence shown above is located. In some embodiments, the shortened protein has a C-terminus where residue 168 (threonine) of the wild-type sequence shown above is located. In some embodiments, the shortened protein has a C-terminus where residues 169, 170, 172, 173, 174, 175, or 176 of the wild-type sequence shown above are located.

[0092] In some embodiments, the abbreviated CutA1 begins with any of residues 30–67 of the human CutA1 sequence shown above. In some embodiments, the abbreviated CutA1 begins with any of residues 44–67. In certain embodiments, the abbreviated CutA1 consists of residues 44–179, residues 61–168, or residues 60–171. In some embodiments, the abbreviated CutA1 begins with any of residues 44–67 and ends with any of residues 168–179. In other embodiments, the abbreviated CutA1 begins with any of residues 30–65 and ends with any of residues 165–179. In some embodiments, the abbreviated CutA1 begins with any of residues 44–67 and ends with any of residues 171–179.

[0093] Shortening CutA1 can improve the accuracy with which fusion structures can be built.

[0094] In certain embodiments, CutA1 is a human CutA1 that has been genetically engineered to have one or more cysteine ​​residues removed from its native sequence, resulting in a shortened N-terminus and / or C-terminus. In some embodiments, the shortened CutA1 consists of residues 44–179 or residues 60–171 and has cysteine ​​substitutions including (i) C75A, C96A or (ii) C75V, C96S. In other embodiments, the shortened CutA1 also begins at any of residues 30–65 and ends at any of residues 165–179 and also has cysteine ​​substitutions including (i) C75A, C96A or (ii) C75V, C96S.

[0095] In certain embodiments, CutA1 is a human CutA1 in which one or more cysteine ​​residues are removed from the native sequence, the N-terminus and / or C-terminus is shortened, and at least one non-cysteine ​​residue is substituted with a cysteine ​​residue. In some embodiments, the shortened CutA1 consists of residues 44-179 of human CutA1 (as shown above) and has cysteine ​​residue substitution removal from the native CutA1 sequence, including or comprising (i) C75A, C96A or (ii) C75V, C96S, and has at least one cysteine ​​located at a residue position that is not a cysteine ​​residue in the native sequence.

[0096] In some embodiments, the shortened CutA1 consists of residues 44-179 of human CutA1 and has cysteine ​​substitution removals including or consisting of C75A and C96A, and cysteine ​​residue substitutions added to one, two, or three of residues K82, E114, and F136.

[0097] In some embodiments, the shortened CutA1 consists of residues 44-179 of human CutA1 and has cysteine ​​substitution removals including or consisting of C75A and C96A, and cysteine ​​residue substitutions added to one or more of residues V64, E78, K79, E83, K91, Q102, K110, S139, F158, and Q166. In some embodiments, this variant CutA1 has cysteine ​​residue substitutions added to 2, 3, 4, 5, 6, 7, 8, 9, or 10 of residues V64, E78, K79, E83, K91, Q102, K110, S139, F158, and Q166.

[0098] In some embodiments, the shortened CutA1 consists of residues 44-179 of human CutA1 and has cysteine ​​substitution removals including or consisting of C75A and C96A, and cysteine ​​residue substitution additions to one or more of residues E78, Q102, and Q166. In some embodiments, this variant CutA1 has cysteine ​​residue substitution additions to two or three of residues E78, Q102, and Q166.

[0099] In some embodiments, the shortened CutA1 consists of residues 44-179 of human CutA1 and has cysteine ​​substitution removals including or consisting of C75A and C96A, and cysteine ​​residue substitutions added to one or more of residues E78, K82, Q102, E114, F136, and Q166. In some embodiments, this variant CutA1 has cysteine ​​residue substitutions added to 2, 3, 4, 5, or 6 of residues E78, K82, Q102, E114, F136, and Q166.

[0100] In some embodiments, variant CutA1 includes or consists of any of the sequences defined above or specific sequences described in the examples, and includes 1 to 10 amino acid substitutions at residues not designated as cysteine ​​residue substitution removal or substitution addition. For example, in some embodiments, variant human CutA1 sequences can be provided that include 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 amino acid substitutions at positions other than 95, 64, 76, 78, 79, 82, 83, 91, 95, 102, 110, 114, 136, 139, 158, and 166. These 1 to 10 substitutions may be substitutions to any of the standard 20 amino acids, but typically, non-cysteine ​​residues will not be replaced with cysteine. Typically, these 1 to 10 substitutions will be conservative substitutions.

[0101] FC Domain The Fc domain of antibodies is well known in this field.

[0102] Typically, the Fc domain (or Fc region) is human Fc, and more typically, human IgG Fc, such as IgG1, IgG2, IgG3, or IgG4.

[0103] The reference human IgG1 (secreted) heavy chain constant sequence has UniProtKB acceptance number P01857-1 (IGHG1). This sequence, as is well known in the art, includes the CH1, hinge (usually defined to include at least CPPCP), CH2, and CH3 regions.

[0104] In some embodiments, the structural domain is typically a human IgG1 constant fragment (Fc) with an integrated hinge region. The hinge region originally contains Cys residues, which enable payload conjugation. In some embodiments, one, two, or three of these cysteine ​​residues can be substituted with non-cysteine ​​residues to control the conjugation.

[0105] For example, the illustrated CC068 sequence (SEQ ID NO: 47) is a SpyCatcher-hinge-Fc-(G4S)2-DogCatcher construct with the C230S (Fc region PDB numbering) mutation. Because there is no flexible linker between the C-terminus of the SpyCatcher3 domain and the hinge region, the C230S mutation was selected to eliminate the N-terminal disulfide bond between the two hinge Fc monomers and to allow for greater conformational freedom in the N-terminal portion of the hinge region. Thus, this construct contains two Cys residues per monomer and four Cys residues per dimer. In contrast, CC60 (SEQ ID NO: 48) results in a SpyCatcher-(G4S)3-Fc-(G4S)3-DogCatcher construct with a deleted hinge region and a longer linker, which can be used as a negative control for payload conjugations.

[0106] Linker as a structural domain In some embodiments, the structural domain includes or consists of polypeptide linkers.

[0107] Polypeptide linkers typically contain or consist of 3 to 30 amino acid residues, typically 5 to 20 residues, for example, 7, 8, 9, 10, 11, 12, 13, or 14 residues. In some embodiments, the polypeptide linker contains or consists of 9 amino acid residues.

[0108] Polypeptide linkers may consist of a large proportion (i.e., more than 50%) of glycine residues and / or serine residues, or may consist of glycine residues and / or serine residues alone. An exemplary linker is GSGGSGGSG.

[0109] The polypeptide linker may be rigid or flexible.

[0110] In some embodiments, the presence of one or more intramolecular disulfide bonds, which help to fix the orientation of the binding domain, provides additional structural support to the binding domain. For example, different domains of a construct can be linked by disulfide linkage. This can be achieved by genetically engineering a non-native cysteine ​​residue at a position in each of the two binding domains that allows for disulfide bond formation between genetically engineered cysteines without disrupting the target binding.

[0111] antigen-binding domain The polyvalent binding polypeptide of the present invention contains multiple binding sites. A typical binding site is the antigen-binding domain, which is well known in the art.

[0112] The antigen-binding domain (sometimes called the antigen-binding region) may be a full-length antibody, an antigen-binding fragment of an antibody, or a genetically modified antibody construct. Antigen-binding domains within the scope of the present invention include, but are not limited to, Fab fragments, F(ab')2, scFv fragments, scFv tandems, scFv-Fc, diabodies, scFv-CH3 (minibodies), scFab, human antibodies, humanized antibodies, nanobodies / VHH fragments, humanized VHH, and (genetically modified / stabilized) human VH domains. Further examples of antigen-binding domains include, but are not limited to, afibodies (genetically modified from protein A), engineered ankyrin repeat proteins (DARPins), or suitable natural or genetically modified protein scaffolds not derived from antibodies, such as Nottin. Furthermore, naturally occurring antigen conjugates such as cytokines and natural ligands or fragments thereof (e.g., growth factors, tumor necrosis factor, interleukins such as TNFα and TGFα; including membrane-bound or soluble components) may be used as antigen-binding domains. As used herein, the antigen-binding domain also refers to receptor fragments that are soluble or capable of binding to membrane-bound proteins or peptides, as exemplified by existing therapeutic drugs such as Enbrel (which comprises an Fc fused to a TNFR2 receptor fragment capable of binding to TNFα and TNFβ).

[0113] In this specification, the term “antibody” is used in its broadest sense and includes certain types of immunoglobulin molecules that contain one or more antigen-binding domains that specifically bind to an antigen or epitope. Antibodies include intact full-length antibodies (e.g., intact immunoglobulins), antibody fragments, and multispecific antibodies.

[0114] A scFv (single-chain variable fragment) typically has a light-chain variable domain (VL) whose C-terminus is connected to the N-terminus of a heavy-chain variable domain (VH) by a polypeptide chain. Alternatively, an scFv may contain a polypeptide chain whose C-terminus of VH is connected to the N-terminus of VL by a polypeptide chain.

[0115] A "Fab fragment" (also called a fragment antigen-binding fragment) contains the constant domain (CL) of the light chain and the first constant domain (CH1) of the heavy chain, along with the variable domains VL and VH of the light chain and heavy chain, respectively. The variable domains contain the complementarity-determining loop (CDR, also called the hypervariable region) involved in antigen binding. The Fab' fragment differs from the Fab fragment in that a few residues are added to the carboxyl terminus of the heavy chain CH1 domain, which contains one or more cysteines in the antibody hinge region.

[0116] The "F(ab')2" fragment contains two Fab' fragments joined by a disulfide bond near the hinge region. The F(ab')2 fragment can be generated, for example, by recombinant methods or by pepsin digestion of an intact antibody. The F(ab') fragment can be dissociated, for example, by treatment with β-mercaptoethanol.

[0117] The "Fv" fragment contains a non-covalently linked dimer of one heavy chain variable domain and one light chain variable domain.

[0118] A "single-chain Fv" or "scFv" comprises the VH and VL domains of an antibody, where these domains reside on a single polypeptide chain. In one embodiment, the Fv polypeptide further includes a polypeptide linker between the VH and VL domains, enabling the scFv to form a structure desirable for antigen binding. For a review of scFv, see Pluckthun in The Pharmacology of Monoclonal Antibodies, vol. 113, Rosenburg and Moore eds., Springer-Verlag, New York, pp. 269-315 (1994). HER2 antibody scFv fragments are described in International Publication No. 93 / 16185; U.S. Patent No. 5,571,894; and U.S. Patent No. 5,587,458.

[0119] The “scFv-Fc” fragment comprises an scFv with an attached Fc domain. For example, the Fc domain may be attached to the C-terminus of the scFv. Depending on the orientation of the variable domain of the scFv (i.e., VH-VL or VL-VH), the Fc domain may follow VH or VL. Any suitable Fc domain known in the art or described herein may be used. In some cases, the Fc domain includes an IgG4 Fc domain. In some cases, the Fc domain includes an IgG1 Fc domain.

[0120] A "single-domain antibody" or "sdAb" refers to a molecule in which one variable domain of the antibody specifically binds to an antigen in the absence of other variable domains. Single-domain antibodies and their fragments are described in Arabi Ghahroudi et al., FEBS Letters, 1998, 414:521-526 and Muyldermans et al., Trends in Biochem. Sci., 2001, 26:230-245. Each of these publications is incorporated by reference in its entirety. Single-domain antibodies are also known as sdAbs or nanobodies. Sdabs are fairly stable and readily expressed as fusion partners with the antibody's Fc chain (Harmsen MM, De Haard HJ (2007). "Properties, production, and applications of camelid single-domain antibody fragments". Appl. Microbiol Biotechnol. 77(1): 13-22).

[0121] In this specification, the terms “full-length antibody,” “intact antibody,” and “whole antibody” are used synonymously and refer to antibodies that have a structure substantially similar to that of naturally occurring antibodies and that have a heavy chain containing an Fc region. For example, when used to refer to an IgG molecule, a “full-length antibody” is an antibody that contains two heavy chains and two light chains.

[0122] Catcher domains and related tags Site-specific isopeptide ligation An isopeptide bond is an amide bond formed between the side chains of two amino acid residues, for example, between the carboxyl group of one amino acid and the amino group of another amino acid.

[0123] Typically, tag / catcher systems that form isopeptide bonds are used according to the present invention. Tag / catcher systems in which alternative chemicals such as ester bonds are used instead of isopeptide bonds are also described.

[0124] The SpyTag / SpyCatcher system is a well-known technique for forming isopeptide links between amino acid residues. Other known tag / catcher systems include SnoopTag / SnoopCatcher, SdyTag / Catcher, and SpyLigase / SpyTag, which can be used in accordance with the present invention.

[0125] SpyTag-SpyCatcher series The SpyCatcher / SpyTag system is known in the art, for example, as described in Zakeri et al. 2012 (Proc Natl Acad Sci US A. 2012 Mar 20; 109(12): E690-E697).

[0126] In short, the Streptococcus pyogenes fibronectin-binding protein FbaB contains a domain (CnaB2) with a spontaneous isopeptide bond between Lys (K31) and Asp (D117). By splitting this domain and rationally genetically engineering the fragments, a peptide (SpyTag) is provided that forms an amide bond with a protein partner (SpyCatcher) in minutes. The catalytic lysine for isopeptide bond formation is present in the SpyCatcher domain, and the reactive aspartic acid is present in the SpyTag peptide. The isopeptide bond formation reaction occurs spontaneously in high yield simply by mixing under a variety of pH, temperature, and buffer / redox conditions. The SpyTag can be fused with a target protein either terminally or internally, and reacts specifically with the SpyCatcher to form a site-directed isopeptide bond. This isopeptide bond has been shown to be irreversible by boiling or peptide competition. Single-molecule dynamic force spectroscopy demonstrated that SpyTag does not separate from SpyCatcher until the covalent bond breaks due to a force exceeding 1 nN. The robust reaction conditions and irreversible ligation of SpyTag highlight spontaneous isopeptide bond formation, providing stable ligation for a novel protein architecture.

[0127] Through multiple rounds of genetic engineering, SpyCatchers / SpyTags with a rapid isopeptide bond formation rate were created. These are being used for protein conjugation with a wide variety of different proteins. For example, the initial SpyTag / SpyCatcher system subsequently evolved from SpyTag / SpyCatcher002 to SpyTag / SpyCatcher003, increasing the isopeptide bond formation rate by 400 times compared to the initial model.

[0128] The range of SpyTag peptides and SpyCatcher polypeptides is known in the art, for example, as described in Arnold et al. (J Am Chem Soc. 2013 September 18; 135(37): 13988-13997), Tirrell et al. (Proc Natl Acad Sci US A. 2014 July 21; 111(31): 11269-11274), Howarth et al. (J Am Chem Soc. 2014 August 21; 136(35): 12355-12363), Li et al. (J Mol Biol. 2014 Jan 23; 426(2): 309-317), and Keeble et al. PNAS December 26, 2019 116 (52) 26523-26533. As will be apparent to those skilled in the art, any preferred arrangement can be used in the present invention.

[0129] Plasmids for basic SpyCatcher and SpyTag constructs are available from the Addgene plasmid repository (www.addgene.org): SpyCatcher (#35044); ΔN1ΔC2 SpyCatcher (#87376); SpyTag-Maltose-Binding Protein (MBP) (#35050); AviTag-SpyCatcher (#72326); SpyCatcher002 (#102827); SpyTag002-MBP (#102831).

[0130] The database of tag / catcher sequences, known as "SpyBank," is publicly available at https: / / www2.bioch.ox.ac.uk / howarth / info.htm and http: / / www.howarthgroup.org / info.

[0131] As of October 2022, this database contained over 1,000 tag / catcher amino acid sequences that could be downloaded as Excel files. SpyBank includes annotated linkers as references for expression organisms, tag / catcher locations (N-terminus, internal, C-terminus), and the basis for creating their fusions. SpyBank itself is described in "Insider information on successful covalent protein coupling with help from SpyBank" by Keeble AH, Howarth M. Meth Enz 2019. As used herein, the terms SpyCatcher and SpyTag refer to the diversity of SC and ST proteins and are not limited to one specific sequence.

[0132] The original SpyCatcher peptide is shown in SEQ ID NO: 2, which has the expressed N-terminal sequence, and SEQ ID NO: 5, which does not have such an optional N-terminal sequence. The original SpyTag peptide is 13 residues long, as shown in SEQ ID NO: 3. A shortened but still functional SpyTag is shown in SEQ ID NO: 4, and a shortened but still functional SpyCatcher is shown in SEQ ID NO: 6. SEQ ID NOs: 7 and 8, and SEQ ID NOs: 8 and 9, each show two further SpyTag / SpyCatcher pairs. The SpyTag in SEQ ID NO: 3 typically pairs with the SpyCatcher in SEQ ID NO: 5 or SEQ ID NO: 6. The SpyTag002 in SEQ ID NO: 7 typically pairs with the SpyCatcher002 in SEQ ID NO: 8. The SpyTag003 (SEQ ID NO: 9) typically pairs with the SpyCatcher003 in SEQ ID NO: 10, but each SpyTag generation can also efficiently bind to other SpyCatcher generations.

[0133] [Table 1-1]

[0134] [Table 1-2]

[0135] Preferred SpyCatcher and SpyTag sequences may include modifications to these exemplary amino acid sequences.

[0136] SpyCatcher variants may include modifications to the amino acid residues of the SpyCatcher peptide. In some embodiments, the SpyCatcher polypeptide contains or consists of sequences that are at least 70% identical, at least 80% identical, at least 90% identical, or at least 95% identical to SEQ ID NO: 2, SEQ ID NO: 5, SEQ ID NO: 6, SEQ ID NO: 8, or SEQ ID NO: 10. In some embodiments, the lysine residues forming the isopeptide bond are not modified. In some embodiments, the lysine residues forming the isopeptide bond are substituted with arginine or histidine.

[0137] A variant of the SpyCatcher molecule usable in this invention is the SpyLigase protein described by Fiereer et al. (PNAS April 1, 2014 111 (13) E1176-E1181). SpyLigase immobilizes two peptide tags together. SpyLigase was formed by splitting CnaB2 into three parts to enable inter-peptide ligation. Specifically, SpyTag(13aa) remained unchanged, while the β-strand of CnaB2 containing reactive Lys was expressed separately and named KTag(10aa). SpyLigase(11kDa) was derived from SpyCatcher by (i) removing residues from the β-strand containing reactive Lys and (ii) circular permutation, replacing the C-terminal residue of CnaB2 with a Gly / Ser linker followed by the N-terminal CnaB2 residue. These sequences are summarized below.

[0138] [ka] Note: TEV protease cleavage site: ENLYFGQ SpyCatcher I>E and M>Y mutations compared to the original CnaB2. SpyTag parent array: AHIVMVDA KTag parent array: ATHIKFSKRD Circular permutations are underlined.

[0139] SpyTag and KTag dock with SpyLigase, and this triplicate structure is arranged similarly to CnaB2 so that SpyTag and KTag directly ligate by covalent bonds. In some embodiments of the present invention, one protein (e.g., an enzyme) is tagged with SpyTag and another protein (e.g., an enzyme) is tagged with KTag. These tagged proteins can be bound together using SpyLigase immobilized on beads.

[0140] In some embodiments, the catcher polypeptide contains or consists of sequences that are at least 70% identical, at least 80% identical, at least 90% identical, or at least 95% identical to the SpyLigase sequences described above.

[0141] SpyTag variants may contain one, two, three, or more modifications to the amino acid residues of the SpyTag peptide. In some embodiments, the SpyTag peptide contains or consists of sequences having one, two, three, four, or five substitutions, additions, or deletions to SEQ ID NO: 3, SEQ ID NO: 7, or SEQ ID NO: 9, with active Asp (e.g., position 7 of SEQ ID NO: 3) remaining unchanged or active aspartic acid being substituted with glutamic acid. Typically, the SpyTag peptide contains the minimal reaction sequence of SEQ ID NO: 4.

[0142] Importantly, catcher / tag systems that are "orthogonal" to SpyCatcher / SpyTag (i.e., do not cross-react with SpyCatcher / SpyTag) are described in the Art. Examples of such orthogonal systems include SnoopCatcher / SnoopTag, SnoopCatcher / SnoopTagJr, and DogCatcher / DogTag. SnoopCatcher and DogCatcher are derived from the same D4 domain of the RrgA adhesion protein of Streptococcus pneumoniae. SnoopCatcher / SnoopTag and DogCatcher / DogTag were formed by differentially splitting the D4 domain at both ends. Because SnoopCatcher and DogCatcher contain congeneral tag sequences, they can react with each other. Based on the RrgA D4 domain, SpyLigase-like systems, namely SnoopLigase and SnoopLigase2 systems with corresponding genealogous tags, have been described.

[0143] Exemplary and non-exclusive catcher-tag combinations are listed in the table below.

[0144] [Table 2]

[0145] First polypeptide One or more domains of the first polypeptide may be linked by a linker, typically a peptide linker. Suitable peptide linkers for use in linking the binding domain to the structural domain are typically 1–150, 1–100, 1–50, 1–25, 1–20, 1–15, or 1–10 amino acid lengths. Examples of linker sequences are GSGS, GGGGS, GGGGSGGGGS, or GGGGSGGGGSGGGGS.

[0146] A number of exemplary first polypeptides are provided below. SpC-Linker1-CutA1-Linker2-SpC SnC-Linker 1-CutA1-Linker 2-SnC SnC-Linker 1-CutA1-Linker 2-SpC SpC-Linker1-CutA1-Linker2-SnC SpC3-Linker1-CutA1-Linker2-DgC Here, SpC is SpyCatcher, SpC3 is SpyCatcher003, SnC is SnoopCatcher, and DgC is DogCatcher. In any case, linkers 1 and 2 are optional. The CutA1 sequence may be human, or derived from Pyrococcus horikoshii, or a homolog from another species, or may have at least 30%, at least 50%, at least 70%, or at least 90% identity with a human or Pyrococcus horikoshii sequence.

[0147] Further examples of structures are as follows: SnC-Linker 1-NC1-Linker 2-SpC SpC-Linker1-NC1-Linker2-SnC In the formula, NC1 is a collagen NC1 domain derived from type VIII or type X collagen. In either case, linker 1 and linker 2 are optional.

[0148] Any suitable structural domain, in particular any of the structural domains described herein, can be used in such structures instead of the exemplary structural domains provided above. Therefore, these exemplary forms are described for use in structural domains in general and in the structural domains described herein.

[0149] Target molecule Examples of target molecules include effector molecules or payloads, which are sometimes referred to in the art, particularly in the context of conjugating a target molecule with an antibody to form an antibody-drug conjugate (ADC). Such payloads and other target molecules are therefore well known in the art. Typically, the target molecule contains or consists of labels, dye molecules, or drugs—substances that, when consumed in the body, exhibit therapeutic or prophylactic physiological effects on the human or animal body. When administered as a drug conjugate, it can enhance or activate the effects of such drugs. Drugs are typically pharmaceuticals.

[0150] In some typical embodiments, the molecule of interest is a drug.

[0151] As will be apparent to those skilled in the art, the drug may be of any type. Drug types that can be conjugated to CutA1 or Fc include, but are not limited to, analgesics, antibiotics, anticancer drugs, anticoagulants, antidepressants, antidiabetics, antiepileptic drugs, antipsychotics, anticonvulsants, antivirals, cardiovascular drugs, depressants, sedatives, and stimulants. In some embodiments, the drug is an anticancer drug, such as a cytotoxic agent. In some embodiments, the anticancer drug is a tubulin inhibitor, such as a metansinoid like meltansine, or auristatin, or a taxol derivative. Examples of drugs for inhibiting tubulin polymerization include auristatins such as tissue factor-directed monomethyl auristatin E (MMAE) and monomethyl auristatin F (MMAF), compounds derived from drastatin 10 (e.g., TZT-1027 described in Kobayashi et al. Jpn J Cancer Res. 1997 Mar;88(3):316-27.), and tubulysins such as tubulysin A. In some embodiments, anticancer drugs are DNA damaging agents that induce cell death by cleaving DNA and / or causing DNA alkylation. Examples of DNA damaging molecules are duocalmycin, calicheamycin, and pyrrolobenzodiazepines. In some embodiments, the drug is a topoisomerase I inhibitor, which binds to the topoisomerase I-DNA complex and stabilizes it, thereby preventing DNA religation and resulting in DNA damage. Examples of drugs that inhibit topoisomerase I are DXd and SN-38. In some embodiments, the drug is an RNA polymerase II inhibitor such as α-amanitin. In some embodiments, the drug is an immunomodulator such as a TLR agonist or STING agonist (see, for example, Table A in Fu et al, Signal Transduction and Targeted Therapy volume 7, Article number: 93 (2022)).

[0152] In some embodiments, the drug is a small molecule drug. Typically, small molecule drugs have a molecular weight of 1000 Da or less, more typically 750 Da or less, or 500 Da or less. This molecular weight is that of the drug molecule itself, excluding any linkers.

[0153] In some embodiments, the drug is a cytotoxic agent, typically a small molecule cytotoxic agent, such as DXd, meltansine, monomethyl auristatin E (MMAE), or monomethyl auristatin F (MMAF). Cytotoxic small molecule drugs are often called chemotherapeutic agents or chemotherapeutic agents, particularly in the context of cancer treatment.

[0154] In some embodiments, the small molecule drug is exatecan.

[0155] Conjugation of a drug with a first polypeptide can be carried out by any known method. Typically, the drug is conjugated to the first polypeptide at a cysteine ​​residue, more typically at a cysteine ​​residue not present in the natural sequence of the structural domain (e.g., CutA1 or Fc domain). When conjugated to cysteine, the molecule to be conjugated may contain a maleimide linker region. The cysteine ​​residue readily reacts with maleimide to form a succinimidyl thioether conjugate.

[0156] In some embodiments, the drug has a molecular weight greater than 1000 Da. In some embodiments, the drug is or comprises an oligopeptide or polypeptide containing two or more amino acid residues covalently linked by peptide bonds, for example, 3, 4, 5, 6, 7, 8, 9, 10 or more, or 20 or more, or 50 or more, or 100 or more.

[0157] In some embodiments, the drug is an immunotoxin, such as Pseudomonas exotoxin A (PE), which is described as an anticancer agent by Wolf and Beile (Int J Med Microbiol. 2009 Mar;299(3):161-76. doi: 10.1016 / j.ijmm.2008.08.003).

[0158] In some embodiments, the molecule of interest is a dye molecule. In some embodiments, the dye molecule is a small molecule dye. Typically, the small molecule dye has a molecular weight of 1000 Da or less, more typically 750 Da or less, or 500 Da or less. In some embodiments, the dye is fluorescent, for example, fluorescein. Conjugation can be carried out by any known method. Typically, the dye is conjugated to CutA1 with a cysteine ​​residue, more typically with a cysteine ​​residue not present in the natural sequence. When conjugated to cysteine, the molecule to be conjugated may contain maleimide. The cysteine ​​residue readily reacts with maleimide to form a succinimidylthioether conjugate.

[0159] In some embodiments, the effector molecule is a nanoparticle, such as a gold nanoparticle, a silica nanoparticle, a lipid nanoparticle, or a lipid polydopamine hybrid nanoparticle ("LPN") as described in Yang et al, Acta Pharmaceutica Sinica B Volume 10, Issue 11, November 2020, pages 2212-2226. Typically, the nanoparticle is conjugated to one or more cysteine ​​residues in CutA1 by site-directed conjugation of the lipid nanoparticle or polydopamine (PDA) hybrid nanoparticle to a maleimide group.

[0160] Post-assembly cleanup In some embodiments, as shown in the examples, the assembly of conjugates (e.g., ADCs, AOCs) provided by the present invention can be combined with a simple post-assembly cleanup. This is particularly useful for the production of drug candidates for downstream analysis. In some embodiments, such methods can be automated. In some embodiments, the cleanup uses beads, such as paramagnetic beads, which can be bound to a suitable catcher to quench unconjugated tagged proteins from the assembly of tagged conjugates and catcher cores (see Figure 17). This enables cost-effective, rapid, and easily scalable purification of assemblies (by removing unbound tagged conjugates).

[0161] A cleanup method typically involves contacting the post-assembly reaction mixture with cleanup binding domains that can bind to unbound conjugates. The cleanup binding domains that have then bound to previously unbound conjugates can then be removed from the reaction mixture.

[0162] These cleanup methods are particularly suitable for plate-based formats, which can be automated using liquid-handling robots. For example, GST catchers can be loaded onto glutathione conjugate beads (e.g., Figure 15), and the unbound excess can be removed by washing with PBS (e.g., Figure 16). The GST catcher-bound beads can then be added to each conjugation reaction in an excess amount, typically relative to the corresponding unconjugated tagged conjugate (e.g., 2 × excess of GST catchers). The beads can then be removed by appropriate means. For example, if the beads are magnetic, they can be removed by magnetic trapping of the particles, leaving a supernatant containing only the fully assembled molecules (e.g., Figure 18). This method is advantageous because it can be implemented on demand at various scales compared to methods that covalently conjugate the catcher to the beads. It also does not require complex chemical processes or reducing conditions. Furthermore, the design of the catcher construct, the ratio of catcher proteins, or the resin can be easily modified according to demand.

[0163] This cleanup technique is a general improvement to the art and may also be applicable to other modular assembly methods, particularly the method described in Driscoll et al, September 2023, bioRxiv 2023.08.31.555700; doi: https: / / doi.org / 10.1101 / 2023.08.31.555700.

[0164] This cleanup technique can be adapted to methods such as sortase-based assembly, as described in Andres et al, Mol Cancer Ther (2020) 19 (4): 1080-1088 "High-Throughput Generation of Bispecific Binding Proteins by Sortase A-Mediated Coupling for Direct Functional Screening in Cell Culture".

[0165] The cleanup step typically uses a fusion protein that can bind to excess conjugates in the reaction mixture after the assembly of the catcher and tagged conjugates. Such cleanup fusion proteins typically comprise a protein useful for the cleanup method described most recently above, fused to at least one binding domain as described herein, typically a catcher domain capable of binding to excess conjugates in the reaction mixture. Proteins useful for the cleanup step can typically form a protein domain-based attachment useful for the purification process. This attachment may be a halotag, which is a modified haloalkane dehalogenase designed to bind covalently (e.g., covalently to a synthetic ligand containing a chloroalkane linker attached to various useful molecules such as fluorescent dyes, affinity handles, or solid surfaces) or non-covalently (e.g., a maltose-binding protein [MBP]). As mentioned above, this construct is typically a glutathione-S-transferase "GST" sequence capable of binding to GSH on beads. In some embodiments, the cleanup fusion protein is attached to a solid support such as beads. In some embodiments, the beads may be magnetic or paramagnetic.

[0166] GST / GSH is a particularly suitable system because relatively low-cost, high-protein-capacity paramagnetic beads are commercially available.

[0167] A cleanup fusion protein may contain a single-type binding domain, thereby enabling the cleanup of one excess conjugate. This single-binding domain may be provided in a single copy (e.g., in GST-SpyCatcher) or, optionally, in multiple copies (e.g., in the form of GST-SpyCatcher-SpyCatcher). If there are two or more excess conjugates after assembly to be cleaned up, two or more cleanup fusion proteins, each having a different single-type binding domain, can be used in combination. GST-SpyCatcher and GST-DogCatcher are examples of cleanup fusion proteins. In some exemplary embodiments, for example, the first cleanup fusion protein is GST-SpyCatcher and the second cleanup fusion protein is GST-DogCatcher. Specific sequences are provided in the examples for GST-SpyCatcher003 and GST-DogCatcher. In other exemplary embodiments, the cleanup fusion protein includes GST-SpyCatcher002 or GST-SpyCatcher-003. In other exemplary embodiments, the cleanup fusion protein includes GST-SpyCatcher002 and GST-SpyCatcher-003 used in combination. Other cleanup fusion proteins include inactive variants of SpyCatcher such as SpyCatcher002KA, SpyDock, or SpySwitch (see, for example, Khairil Anuar et al, Nature Communications volume 10, Article number: 1734 (2019), particularly Figure 7).

[0168] MBP-SpyCatcher and MBP-DogCatcher are further examples of cleanup fusion proteins. In some exemplary embodiments, for example, the first cleanup fusion protein is MBP-SpyCatcher and the second cleanup fusion protein is MBP-DogCatcher. In other exemplary embodiments, the cleanup fusion protein includes MBP-SpyCatcher002 or MBP-SpyCatcher-003. In other exemplary embodiments, the cleanup fusion protein includes MBP-SpyCatcher002 and MBP-SpyCatcher-003 used in combination.

[0169] Halotag-SpyCatcher and Halotag-DogCatcher are further examples of cleanup fusion proteins. In some exemplary embodiments, for example, the first cleanup fusion protein is Halotag-SpyCatcher and the second cleanup fusion protein is Halotag-DogCatcher. In other exemplary embodiments, the cleanup fusion protein includes Halotag-SpyCatcher002 or Halotag-SpyCatcher-003. In other exemplary embodiments, the cleanup fusion protein includes Halotag-SpyCatcher002 and Halotag-SpyCatcher-003 used in combination.

[0170] In some embodiments, two or more binding domains are attached to a single protein useful for the cleanup method described above. Multiple binding domains, typically multiple catcher domains, can be attached to any preferred location on the protein useful for cleanup (e.g., GST, MBP, halotag). Multiple binding (e.g., catcher) domains may be arranged in series at one end of the cleanup protein, for example, or at different ends. As with other embodiments described herein, a linker peptide (e.g., 2 to 30 amino acid residues) may be included between the separate components of the fusion as needed. Some exemplary configurations where each catcher differs include Catcher1-GST-Catcher2, Catcher1-Catcher2-GST, GST-Catcher1-Catcher2, Catcher1-MBP-Catcher2, Catcher1-Catcher2-MBP, MBP-Catcher1-Catcher2, Catcher1-halotag-Catcher2, Catcher1-Catcher2-halotag, or halotag-Catcher1-Catcher2.

[0171] In some embodiments, the useful polypeptide comprises a first binding domain at the N-terminus and a second binding domain at the C-terminus, wherein the first and second binding domains are separated by a structural domain useful for the cleanup method described most recently above, and the first and second binding domains are either the same or different. In such embodiments, the structural domain can typically form a protein domain-based attachment useful for the purification process. This attachment may be covalent (e.g., a halotag) or non-covalent (e.g., a maltose-binding protein [MBP]). As mentioned above, this construct is typically a glutathione-S-transferase "GST" sequence that can bind to GSH on beads.

[0172] The present invention will be further described with reference to the following non-limiting embodiments. [Examples]

[0173] The inventors have devised a method for the large-scale production of bispecific molecules (or other-order multispecific molecules, such as triplicate molecules) characterized by conjugations, such as small molecule dye conjugations or small molecule drug conjugations. This method is described in Example 1 and Figure 1.

[0174] Example 1: Design of a scalable process for generating conjugate bispecific drug candidates Modular assembly systems based on SpyCatcher / SpyTag and DogCatcher / DogTag chemistry enable the rapid generation of multivalent bispecific drug candidates. In addition, this system allows for the large-scale generation of multivalent bispecific payload conjugate drug candidates by introducing site-directed mutations into either the core protein (such as CutA1 or Fc) or the SpyCatcher / DogCatcher domain, thereby covalently linking the catcher-core protein to a fluorescent or cytotoxic small molecule (Figure 1). First, the catcher-core protein is bulk conjugated to the target small molecule (via the attachment site of the target molecule), then purified, and subsequently assembled with SpyTag / DogTag ligand proteins to generate a library of bispecific small molecule conjugate drug candidates. Using the example shown in Figure 1, which generates a bispecificity ADC library of size n using a single drug-linker compound, conventional methods require n reactions to generate n bispecificity candidates, followed by another n reactions to conjugate each bispecificity candidate with the drug-linker compound, each potentially requiring further processes such as post-conjugation dialysis and cleanup. In contrast to conventional ADC library generation methods, our method requires payload conjugation and payload-protein conjugate purification to be performed only once, rather than for each individual drug candidate in the library. This is advantageous because protein-payload conjugate purification can be complex and may lead to the loss of approximately 40-50% of the original sample. In our method, sample loss is limited to the catcher-core only. The conjugation reaction can be scaled up to the volume required for specific ADC library screening. Therefore, catcher-core proteins conjugated to different payloads can be stored at -80°C, then thawed as needed, assembled with conjugates (such as antigen-binding domains), and used to generate an ADC library.This enables parallel, large-scale screening of bispecific drug candidates with different toxic payloads.

[0175] Example 2: Rational CutA1 core protein gene manipulation for terminal shortening and native cysteine ​​removal CutA1 possesses several advantageous characteristics for use as a core protein in bispecific molecules. Nevertheless, modifying the protein to enhance its compatibility with workflows described elsewhere herein may be beneficial. Such modifications include shortening the terminus of HsCutA1 to remove regions not resolved in the crystal structure (PDB ID: 2ZFH), forcing similar distances between the N-terminal and C-terminal core residues and the core surface, and removing native unpaired cysteine ​​from the HsCutA1 sequence due to potential interference with downstream processes.

[0176] Other than in this specification, full-length HsCutA1 (SEQ ID NO: 1), HsCutA1 containing signal peptides are mentioned. 44~179 To obtain this, 43 N-terminal residues were removed (SEQ ID NOs: 15, 16, 17), and HsCutA1 60~171 A version obtained by removing 59 N-terminal residues and 11 C-terminal residues (SEQ ID NO: 18) is described. HsCutA1 60~171 By removing 59 N-terminal residues and 8 C-terminal residues to obtain (SEQ ID NO: 19), the distance of the residues protruding from the HsCutA1 core can be made similar (Figure 2). Protein production did not show dramatic differences in expression between the two molecules, suggesting that shortening does not significantly affect HsCutA1 (Figure 2). The cytotoxicity trends of Colo205 cells assembled with tagged L7 were nearly identical between the two molecules, suggesting that the conferred function is unaffected (Figure 2).

[0177] HsCutA1 has two native unpaired cysteines at positions 75 and 96 (Figure 3). Alanine scanning mutagenesis is a common strategy for substituting unpaired cysteine ​​residues. Furthermore, multiple sequence alignments were created using the HsCutA1 sequence to identify preferred amino acid frequencies at positions C75 and C96 across all 406 homologs. Subsequently, these residues were mutated to alanine at both positions to obtain the [C75A,C96A]HsCutA1 mutant, or mutated to valine at position 75 and serine at position 96 to obtain [C75V,C96S]HsCutA1. Protein production did not show dramatic differences in expression between molecules, suggesting that cysteine ​​removal does not significantly affect HsCutA1 (Figure 3). The cytotoxic effects in Colo205 cells when assembled with tagged L7 were nearly equivalent. This suggests that the granted functionality is unaffected (Figure 3).

[0178] Example 3: Mutagenesis of CutA1 as the core protein of a multivalent bispecific drug conjugate To enable site-specific conjugation of the payload to the CutA1 catcher-core, six amino acid positions in CutA1[44~179, C75A, C96A](SEQ ID NO: 20) were mutated to Cys based on homology modeling or surface exposure (see Methods: Cysteine ​​Mutant Design). The thiol groups of the Cys residues could then be reacted with the maleimide moiety of the linker-payload compound (Figure 4). The locations of the mutations in the CutA1 structure are shown in Figure 5A. A comparison of conjugation efficiency between various mutants, such as the CutA1 catcher-core constructs (SEQ ID NOs: 37, 38, 39, 40, 41, 42), as well as the negative control CC041 (SEQ ID NO: 36) and the wild type (SEQ ID NO: 15), is shown in Figure 6A (dye / protein ratio). All newly introduced Cys mutations showed higher conjugation efficiency compared to the wild type and negative control. The mass spectra (Figure 7) of CC042 (SEQ ID NO: 38) conjugated with fluorescein-5-maleimide, deruxtecan, maleimidocaproyl-Val-Cit-DM1, and maleimidocaproyl-Val-Cit-PAB-MMAF are consistent with complete conjugation of a single linker-payload compound with each CutA1 monomer. Since CutA1 is a trimer, this corresponds to a 3:1 “drug-versus-antibody” ratio. Figure 8 shows both the conjugation with fluorescein-5-maleimide, as well as the assembly of CC042 (SEQ ID NO: 38) with SpyTag-tagged and DogTag-tagged ligands. In contrast, the negative control CC041 (SEQ ID NO: 36) presents low levels of nonspecific labeling but assembles with ligands. Figure 8 suggests that maleimide conjugation does not impair the ability of catcher-core proteins to assemble with tagged ligands.

[0179] Example 4: Fc implementation as a core protein for polyvalent bispecific drug conjugation A constant fragment (Fc) of IgG1 incorporating the hinge region can also be used as a core protein. The SpyCatcher domain and the DogCatcher domain can be fused to either end of the hinge Fc to generate the SpyCatcher-hinge-Fc-(G4S)2-DogCatcher protein, or the DogCatcher-SpyCatcher-hinge-Fc protein can be generated as a tandem fusion. The hinge region naturally contains three surface-exposed Cys residues, which enable payload conjugation. In addition, DAR can be controlled by inducing mutations in the Cys residues to Ser residues. For example, CC068 (SEQ ID NO: 47) is a SpyCatcher-hinge-Fc-(G4S)2-DogCatcher construct with the C230S (Fc PDB numbering) mutation. Because there is no flexible linker between the C-terminus of the SpyCatcher3 domain and the hinge region, the C230S mutation was selected to eliminate the N-terminal disulfide bond between the two hinge-Fc monomers and to allow for greater conformational freedom in the N-terminal portion of the hinge region. Thus, this construct contains two surface-exposed Cys residues per monomer and four surface-exposed Cys residues per dimer, and therefore the maximum DAR (drug-to-antibody ratio) is 4. In contrast, CC60 (SEQ ID NO: 48) has a deleted hinge region and uses a longer linker, resulting in the SpyCatcher-(G4S)3-Fc-(G4S)3-DogCatcher construct. Figure 9 shows the conjugation with fluorescein-5-maleimide, as well as the subsequent assembly of CC068 with SpyTag ligand and DogTag ligand. However, in the case of CC060, only the assembly is shown because the hinge region is deleted. Therefore, Fc-based catcher cores that retain the natural hinge Cys residue can also be used to generate a multivalent, bispecific payload conjugate drug candidate library.

[0180] Example 5: Production of a dye-conjugate bispecific substance for rapid imaging without antibody-based staining. Figure 8C illustrates the fluorescein conjugate CutA1 core, which can be assembled with one or two conjugates to produce multivalent and multivalent bispecific fluorescein-labeled assemblies. This system enables the large-scale generation of libraries of multivalent bispecific fluorescein-labeled drug conjugates. These assemblies can then be used for imaging without further staining using fluorescent antibodies (Figure 10C). In addition, this system can be used to screen highly binding / internal migration conjugate clones using high-throughput imaging applications. Because the fluorescein conjugation is uniform throughout the library, binding and internal migration profiles can be quantified across the entire library.

[0181] Example 6: Scalable production of drug conjugate bispecific substances for rapid screening of drug candidates Drug conjugation with catcher-core proteins enables the subsequent generation of libraries of drug-conjugate bispecific substances. These drug conjugates can then be screened for efficacy and specificity using cell viability screening techniques such as cell titer growth assays or high-throughput in vivo imaging based on cell counting. Figures 11–14 illustrate four libraries of drug-conjugate bispecific substances, each containing CC042-DM1, CC068-deruxtecan, CC042-MMAF, and CC068-MMAF, all orthogonally conjugated to conjugates L2, L4, L5, and L6, respectively. Cell viability screening at three concentrations using MTT assays as readings for cell viability of drug-conjugate assemblies and assemblies without drug conjugation rapidly informs about polyvalent and / or bispecific conjugate pairs that benefit from the drug conjugate and improve efficacy. Furthermore, since the DAR of drug conjugation with catcher-core and Fc is controlled, comparing libraries of catcher-core drug conjugates with various DARs / linkers can provide further information about the most promising hits for drug conjugate generation.

[0182] Example 7: By integrating modular assembly with simple post-assembly cleanup, it becomes possible to produce uniform drug candidates for downstream analysis. Using commercially available paramagnetic beads, GST-SpyCatcher003 ("GST-SpC3", SEQ ID NO: 49) and GST-DogCatcher ("GST-DgC", SEQ ID NO: 50), which are conjugated to commercially available paramagnetic beads to quench unconjugated tagged proteins from assemblies of SpT and DgT conjugates and catcher-cores, an automated method was developed (Figure 17). This enables cost-effective and easily scalable rapid purification of assemblies. This is suitable for a low-volume plate-based format that can be automated using a liquid-handling robot. GST-catchers are loaded onto beads (Figure 15), and any unconjugated excess is washed off with PBS (Figure 16). GST-catcher-conjugated beads are added to each conjugation reaction to obtain a 2× excess of GST-catcher relative to the corresponding unconjugated L2-DgT and L7-SpT. Next, these can be removed by magnetic trapping of the particles, leaving a supernatant containing only the fully assembled molecules (e.g., Figure 18). This method is advantageous because it can be implemented on demand at various scales compared to methods that covalently conjugate catchers to beads. Furthermore, it does not require complex chemical processes or reducing conditions. In addition, the design of the catcher construct, the ratio of catcher proteins, or the resin can be easily modified according to the demands.

[0183] Example 8: A large modular assembly can be used to rapidly screen a wide range of panel target combinations in a 384-well plate format. To further illustrate the potential for scaling up this method, the inventors demonstrated the implementation of a higher-throughput robotic method during drug candidate generation, as well as the miniaturization of drug administration into 384-well plate assemblies. Figures 19–21 illustrate the cytotoxicity after drug administration of CC068 conjugated with MMAF with two linker variants (with and without PEG) to cell lines at three concentrations, and a comparison with a panel of drug candidates derived solely from CC068.

[0184] In the assays illustrated in Figures 19–21, DgT conjugates were added to non-mammalian targets to generate functionally monospecific controls of the SpT conjugates, more closely matching the process of bispecific drug candidates, and an alternative set of monospecific controls is provided in Figures 11–14. In Figure 22, this method was extended using additional corresponding SpT conjugates to non-mammalian targets for monospecific controls derived from the DgT conjugates. A comparison between Figures 20 and 21 and Figure 19 demonstrates that cytotoxicity is highly payload-dependent, as well as target-dependent payload delivery. Despite having the same payload, the differences between MMAF and PEG-MMAF conjugates further highlight the applicability of the method to testing various aspects of drug design. In addition to Figures 19–21, the inventors performed a similar screening derived from CC042, which further highlights the influence of core geometric shape on drug conjugate behavior. In summary, the inventors demonstrated that this method can rapidly derive drug candidates with different linker-payload conjugations from a given pre-conjugated protein, and that it can be applied to a wide range of conjugate families.

[0185] Finally, Figure 22 shows that this method is suitable for accelerating the entire process and conducting screening campaigns on a commercially competitive scale. To demonstrate the capabilities of this method, the inventors further implemented a robotic system for high-throughput application of drug candidates to target cell lines. Hereinafter, the inventors were able to generate a panel of approximately 800 drug candidates, of which approximately 400 were drug conjugates conjugated to PEG-MMAE (344 bispecific drug conjugates + 8 monospecific tetravalent assemblies [SpT and DgT conjugates are identical] + 52 having a second conjugate to a non-mammalian target as a functionally monospecific drug candidate), which could proceed to drug administration after <1 week. In total, a panel of 9 SpT conjugates and 45 DgT conjugates was used (of which one SpT conjugate and one DgT conjugate had a target unrelated to mammalian cells as a “monospecific” control). This covered a wide range of different targets (5 for SpT and 28 for DgT), facilitating the identification of novel target combinations. Similar to Figures 19–21, Figure 22 illustrates the impact of payload conjugation and target selection, as well as the differences between bispecific and monospecific assemblies, as would be expected from the BsADC screening panel. Figure 23 further analyzes these differences between bispecific and related monospecific assemblies, and also analyzes the changes between drug-conjugated and non-conjugated assemblies, along with the changes in overall cytotoxicity, allowing for the clustering of bispecific drug-conjugated assemblies to gain information about candidate selection.

[0186] Example 9: Catcher-core designs with various geometric shapes, drug-to-antibody ratios (DARs), and other modifications can be derived from the same core molecule. In addition to catcher domains separated by a single core molecule, such as SpyCatcher003-Fc-DogCatcher of CC068 having SEQ ID NO: 47, we have also derived alternative constructs from Fc, including SpyCatcher003-DogCatcher-Fc of CC076 having SEQ ID NO: 52 (illustrated in Figures 1 and 24). Such platforms with alternative geometric shapes provide a useful method for evaluating the influence of conjugate configuration or valency on drug candidate behavior (comparisons between different core proteins are also shown in Figures 11–14).

[0187] Figure 24 shows that CC076 is suitable for assembly with SpT and DgT proteins, providing a novel protein platform useful for this and related screening methods (e.g., conjugation-free bispecificity screening, similar to International Publication No. 2022 / 200804). CC076 is also suitable for pre-conjugation to small molecules (such as drug conjugations as shown herein by fluorophore conjugation) followed by SpT / DgT-based protein assembly. This form is therefore a suitable catcher-core (CC) protein in the context of the present invention. Figure 25 provides a drug screening based on a set of bispecificity drug candidates assembled to CC068 and CC076 without payloads. This illustrates the differences in cytotoxicity due to changes in geometric shape.

[0188] Similarly, SpC-Fc / DgC-Fc complexes for subsequent assembly can be created using knob-into-hole or related Fc heterodimerization techniques. Herein, pre-Fc heterodimerization payload conjugations can be used to create pre-conjugate catcher-core constructs with reduced DAR (e.g., a pre-hybridization single SpC-Fc or DgC-Fc payload conjugation resulting in a conjugated Fc hybrid of one half of the construct) or two different linker-payload species (e.g., conjugations of different linker-payload species with pre-hybridization SpC-Fc and DgC-Fc). Similarly, post-Fc heterodimerization payload conjugates are also suitable for providing assembly-ready pre-conjugate complexes. The inventors have recognized that, for example, when it is desired to maintain the same positional relationship between the conjugate and the protein tag relative to the Fc, knob-into-hole type Fc heterodimers, such as SpC-Fc / SpC-TEV-Fc, may be preferred for certain protein-protein assembly methods or related techniques for the sequential assembly of SpT conjugates.

[0189] In addition to deriving alternative drug candidate geometric shapes, the core protein can also be modified to include useful mutations or modifications such as “silencing” Fc mutations known in the art (e.g., “LALA”, “LAGA”, “LALAPG”), or to introduce or remove cysteine ​​residues or other residues for linker-payload attachment, thereby enabling the generation of catcher core molecules with various DARs. For example, in in vitro assays, it may be desirable to evaluate silencing Fc-based assemblies instead of or compared to wild-type Fc-based assemblies in order to eliminate the potential influence of the Fcγ receptor on the internal transfer of drug conjugates. We show that additional cysteine ​​to provide more small molecule conjugation sites in catcher-core variants, e.g., variations of CC068 having eight Cys residues per dimer of the Fc-based catcher-core facilitate drug-to-antibody ratios of up to 8. Such molecules are CC080 (SEQ ID NO: 53) and CC086 (SEQ ID NO: 54).

[0190] In general, the “conjugation-first” methods described herein can utilize a wide variety of techniques, either sequentially or in addition to any single step, to generate a broad range of bispecificity, multispecificity, geometric shapes, small molecule conjugations, combinations of small molecule conjugations, or other features.

[0191] Example 10: Information on drug design of bispecific antibody-drug conjugates. The screening methods described herein are useful for identifying novel targets, target combinations, and / or conjugate combinations for payload delivery. A wide range of target combinations were tested in Figure 22.

[0192] The results of these tests indicate that several such target combinations are of interest in the field of bispecific ADC technology. Furthermore, Figure 22 shows that several identified target combinations can be repurposed between different conjugates for the same target (e.g., target 2) and between different geometric shapes (i.e., SpT conjugate A + DgT conjugate B compared to SpT conjugate B + DgT conjugate A in the context of an assembly derived from SpC-Fc-DgC; see the left quadrant of Figure 22). This provides evidence of the internal consistency of this discovery method.

[0193] In this specification, variations in repeated screening triage and drug candidate design provide useful indicators for designing the final drug candidate. At any stage, if an interesting conjugate or target combination is obtained, the drug candidate can be further optimized to improve its performance. From this initial discovery phase, various methods for generating bsADCs in forms suitable for drug production, such as DuoBody, CrossMAb, FIT-Ig, DVD-Ig, Morrison forms, scFv-IgG, or related technologies, are known in the art. Barron et al., 2024 (https: / / www.ncbi.nlm.nih.gov / pmc / articles / PMC10889805 / ) and International Publication No. 2022 / 200804 describe comparisons between bispecific candidates or bispecific ADCs at the screening stage and drug-like forms (e.g., changes in DAR or removal of adapter-binding domains during conversion to the final molecule). Similarly, the usefulness of “conjugate-related” methods for the discovery of bsADCs is well known in the art, and commercial use is available through anti-Fc Fab conjugates (which can be combined with bispecific antibodies to produce complexes that mimic bispecific ADCs; for example, “anti-HIgG(Fc)Fab-C-MMAE ADC (catalog number: ADC-1011-SHFz)” sold by Creative Biolabs [as of June 13, 2024, https: / / www.creativebiolabs.net / hFc-Fab-C-MMAE-22302.htm]). While we do not wish to be bound by theory, we believe that the methods provided herein can be used to identify previously unknown target combinations from scales and variations that were previously difficult or impossible to achieve. Once a given combination is selected for further development, methods known in the art are suitable for deriving drug molecules that reproduce or surpass the newly discovered effects.

[0194] Figure 26 illustrates, as an example, drug-like bispecific ADCs corresponding to target combinations that showed increased cytotoxicity in the screening shown in part of Figure 22. Figure 26 shows that drug-like bispecific ADCs with favorable binding properties and high cytotoxicity as MMAE drug conjugates can result in reduced cell viability, as also shown in the bispecific drug conjugate assembly for the same target combination in Figure 22 (in this case, approximately 40% at a 2.5 nM dimer concentration for the same cell line as described in Figure 26).

[0195] method Multiple array alignment: Cys mutations were selected using multi-sequence alignment. Sequence ID 1 was used as the query for the HMMER (Eddy, 1998, Bioinformatics) search in the UniProtKB database (v.2021_04) (The UniProt Consortium, 2020, Nucleic Acids Research). An e-value cutoff of 0.0001 was used. Sequences with alignment lengths shorter than 60% of Sequence ID 1 were discarded. Redundant sequences were then removed using CD-HIT (Li and Godzik, 2006, Bioinformatics) with a sequence identity threshold of 0.95. Multi-sequence alignment was then performed using MAFFT (Katoh et al, 2002, Nucleic Acids Research). The multi-sequence alignments obtained from MAFFT were analyzed using the Biopython, NumPy, and Pandas libraries in Python 3 to calculate the frequency of each amino acid at each position in Sequence ID 1. Gaps were excluded from this analysis.

[0196] Cysteine ​​variant design Cysteine ​​mutants K82C, E114C, F136C The human CutA1 crystal structure (PDB ID 2ZFH) was aligned to its homologous structure using FATCAT (Ye, Y. & Godzik, A. FATCAT: a web server for flexible structure comparison and structure similarity searching. Nucleic Acids Res. 32, W582-585 (2004)). The corresponding surface positions in the homolog opposite the Cys residue were selected, leading to the following mutations: K82C, E114C, F136C (SEQ ID NOs. 31, 33, 34).

[0197] Cysteine ​​mutants E78C, Q102C, Q166C The human CutA1 crystal structure (PDB ID 2ZFH) was minimized using the RepairStructure command with the EvoEF2 force field (Huang, X., Pearce, R. & Zhang, Y. EvoEF2: accurate and fast energy function for computational protein design. Bioinforma. Oxf. Engl. 36, 1135-1142 (2020)). Then, the BuildMutant command was used to mutate Cys75 and Cys96 to Ala. Subsequently, this model with Cys-to-Ala mutations was used to generate structural models with single Cys mutations at positions V64, E78, K79, E83, K91, Q102, K110, S139, F158, and Q166. The solvent exposure of in silico Cys mutant models was evaluated by averaging the solvent exposure area of ​​mutant residues in three strands of CutA1 using PDBePISA (Krissinel, E. & Henrick, K. Inference of Macromolecular Assemblies from Crystalline State. J. Mol. Biol. 372, 774-797 (2007)). Three mutants were selected across a range of solvent exposure levels: E78C, Q102C, and Q166C (SEQ ID NOs. 30, 32, and 35).

[0198] [Table 3]

[0199] All of the above Cys mutations were individually introduced into the SpC3-G4S-HsCutA1[44~179, C75A, C96A]-G4S-DgC construct (SEQ ID NO: 36).

[0200] Molecular cloning Plasmids encoding recombinant proteins were provided by Twist Biosciences or ProteoGenix. DNA fragments and oligonucleotides were synthesized by Integrated DNA Technologies (IDT). Constructs for SpC3-Hinge-Fc-DgC, SpC3-Fc-DgC, and SpC3-HsCutA1-DgC were assembled using standard cloning procedures. Standard polymerase chain reaction (PCR) was used to amplify the synthesized DNA fragments to introduce point mutations into the plasmid backbone, or to make other adjustments to the recombinant sequence, followed by standard cloning methods such as restriction cloning. The assembled constructs were transformed into E. coli NEB5-alpha cells. Putative-positive clones were grown overnight, and DNA was isolated from the bacterial pellet by miniprep. Samples were sent to Source Bioscience for Sanger sequencing for sequence validation, and alignment was performed using Benchling's molecular biology package (www.benchling.com).

[0201] Protein expression and purification Expression To obtain proteins CC7, CC041, CC038, CC042, CC044, CC056, CC057, CC058, and L2, L4, L5, L6, L7, and L8, DNA encoding these genes in a pET28 expression vector was used to transform BL21(DE3) or SHuffle T7 expressing cells. Colonies were inoculated into LB cultures containing 50 μg / mL kanamycin with shaking at 160–220 rpm at 37°C. The cultures were diluted 1:100 overnight with LB or 2×YT medium supplemented with 50 μg / mL kanamycin. The cultures were grown at 37°C with shaking at 160–220 rpm, and protein expression was induced with 0.2–0.4 μM IPTG at OD 0.6–0.8 (LB) or 1.6–2.0 (2×YT). The cultures were further incubated at 37°C for 4 hours or at 18°C ​​for 16 hours while shaking at 160-200 rpm.

[0202] Alternatively, the culture was diluted 1:100 overnight in SB Automated Induction Medium supplemented with 50 μg / mL kanamycin. The culture was grown at 37°C for 4 hours with shaking at 160–220 rpm. Then, the culture was incubated at 30°C for a further 16 hours with shaking at 160–200 rpm.

[0203] Cells were collected by centrifugation at 5000 × g for 15 minutes at 4°C, and the pellet was stored at -80°C before purification.

[0204] To obtain proteins CC060 and CC068, the DNA encoding these genes from the pTWIST expression vector was transfected into Expi293F cells (ThermoFisher) in Expi293 expression medium using PEI MAX reagent (Polysciences) and Opti-MEM reduced serum medium. The cells were incubated at 37°C and 8% CO2 for 16 hours with shaking at 120 rpm. Subsequently, the cells were supplemented with a sterile (using a 0.22 μm syringe filter) phyton / sodium butyrate expression adjuvant to final concentrations of 0.5% phyton (Appleton Woods) and 2 mM sodium butyrate (ThermoFisher). The cells were further incubated at 37°C and 8% CO2 for 4–6 days with shaking at 120 rpm. To obtain protein CC076, the DNA encoding these genes from the pTWIST expression vector was transfected into Expi293F cells as described above. The proteins CC080, CC086, the conjugates shown in Figures 19-22, and the proteins bAb001 and bAb003 were either expressed as described above and purified as described below, or produced by other means known in the art, such as expression from ExpiCHO, Expi293, or CHO-K1 cells followed by purification with protein A or Ni-NTA.

[0205] Protein A The mammalian Expi293F supernatants of CC060, CC076, and CC068 were collected by centrifugation at 300×g for 30 minutes. The supernatants were supplemented with 1 mM PMSF and a 1× cOmplete EDTA-free protease inhibitor cocktail and filtered.

[0206] The supernatant proteins were purified using AKTA Pure FPLC with a HiTrap MAbSelect PrismA column (Cytiva). The samples were purified at 4°C using 20 mM sodium phosphate, 150 mM NaCl, pH 7.2 start buffer, and 0.1 M glycine, pH 3.0 elution buffer. 1 M Tris-HCl, pH 8.0 neutralization buffer was added to the wells of a 96-well deep-well collection plate. Elution peak fractions were analyzed by SDS-PAGE, and appropriate fractions were pooled. For larger volumes, these proteins were further purified using a HiScreen Fibro PrismA column (Cytiva).

[0207] Ni-NTA For proteins CC038, CC042, CC044, CC056, CC057, and CC058, bacterial cell pellets were resuspended in Ni-NTA equilibrium buffer (50 mM Tris, pH 7.8, 300 mM 25NaCl, 10 mM imidazole) supplemented with 1 mM PMSF, a 1×cOmplete EDTA protease inhibitor cocktail, benzonase (5 U / mL), 1 mg / mL lysozyme, and 5 mM 2-mercaptoethanol. Samples were sonicated for 9–12 minutes using an ultrasonic processor with a 20 mm probe and 20% amplitude, with pulses of 2 seconds on and 4 seconds off.

[0208] Alternatively, bacterial cell pellets were resuspended in B-PER chemilysis buffer (ThermoFisher) supplemented with 1 mM PMSF, 1 × cOmplete EDTA-free protease inhibitor cocktail, benzonase (5 U / mL), 1 μg / mL lysozyme, and 5 mM 2-mercaptoethanol.

[0209] Next, the sample was centrifuged at 16,000 × g for 30 minutes. The supernatant was retained for Ni-NTA chromatography.

[0210] The supernatant was loaded onto a pre-equilibriumized HisPur Ni-NTA gravity flow column. The resin was washed with Ni-NTA wash buffer 1 (50 mM Tris, pH 7.8, 300 mM NaCl, 10 mM imidazole) supplemented with 5 mM 2-mercaptoethanol at a volume 20 times the volume of the resin bed, and then washed again with Ni-NTA wash buffer 2 (50 mM Tris, pH 7.8, 300 mM NaCl, 30 mM imidazole) supplemented with 5 mM 2-mercaptoethanol at a volume 10 times the volume of the resin bed. His-tagged proteins were eluted from the resin with Ni-NTA elution buffer (50 mM Tris, pH 7.8; 300 mM NaCl, 200 mM imidazole) supplemented with 5 mM 2-mercaptoethanol at a volume 5 times the volume of the resin bed until the absorbance of the eluted fraction at 280 nm approached the baseline.

[0211] For proteins CC7, CC041, L2, L4, L5, L6, L7, and L8, the same protocol was used, but the buffer was not supplemented with 5 mM 2-mercaptoethanol.

[0212] SEC For CC038, CC042, CC044, CC056, CC057, and CC058, proteins with a purity exceeding 90% after Ni-NTA chromatography were dialyzed against PBS supplemented with 5 mM 2-mercaptoethanol using SnakeSkin® dialysis tubing from 3K MWCO. The proteins were then concentrated in Pierce® Protein Concentrator PES. The proteins were then further purified using an AKTA Pure FPLC with a Superdex 16 / 600 S200 column. The samples were purified at 4°C using PBS buffer supplemented with 2 mM TCEP.

[0213] For CC7, CC041, L2, L4, L5, L6, L7, and L8, the same protocol was used, but the buffer was not supplemented with 2-mercaptoethanol and TCEP.

[0214] For CC060, CC076, and CC068, optional dialysis was performed on PBS using 3K MWCO SnakeSkin® dialysis tubing after protein A purification, followed by concentration and purification of the buffer as described above, without supplementation with 2-mercaptoethanol and TCEP.

[0215] The elution peak fractions were analyzed by SDS-PAGE, and appropriate fractions were pooled. Absorbance at 280 nm was measured using an Implen NanoPhotometer N60, and the estimated protein concentration was calculated using the reduction extinction coefficient predicted by ProtParam. Protein samples were adjusted to monomer concentrations of 10–300 μM and stored at -80°C.

[0216] Cysteine-maleimide conjugation Catcher-core protein reduction Catcher-core protein variants (monomers) containing a single Cys residue were prepared to a concentration of 50 μM (monomer) in PBS buffer pH 7.4 (Formedium) with 50 mM EDTA (Formedium). To reduce the cysteine ​​residue, 100 molar equivalents (relative to the protein thiol) of 40 mM TCEP (Melford) in PBS buffer pH 7.0 were added to the catcher-core protein, and the mixture was incubated in a microplate shaker (Fisher Scientific) at 600 rpm for either a) 16 hours at 37°C or b) 2 hours at 25°C.

[0217] Alternatively, catcher-core protein variants with more than one Cys residue per monomer were prepared to a 50 μM monomer concentration in PBS buffer pH 7.4 (Formedium) with 10 or 50 mM EDTA (Formedium). To reduce the cysteine ​​residues, 4–25 molar equivalents (relative to protein thiols) of TCEP (Melford) in PBS buffer pH 7.0 were added to the catcher-core protein, and the mixture was incubated at 25°C for 2 hours with shaking at 600 rpm in a microplate shaker (Fisher Scientific).

[0218] Bispecific antibody reduction Bispecific antibodies were prepared to a 50 μM dimer concentration in PBS buffer pH 7.4 (Formedium) containing 10 mM EDTA (Formedium). To reduce cysteine ​​residues, 4 or 32 molar equivalents (relative to protein thiols) of TCEP (Melford) in PBS buffer pH 7.0 were added to the catcher-core protein, and the mixture was incubated at 25°C for 2 hours with shaking at 600 rpm in a microplate shaker (Fisher Scientific).

[0219] Maleimide conjugation 10-15 molar equivalents (relative to protein thiols) of fluorescein-5-maleimide, deruxtecan, and MC-Val-Cit-PAB-DM1 or MC-Val-Cit-PAB-MMAF (dissolved in MedChemExpress or DMSO) were added, and the mixture was incubated at 25°C for 1 hour. Finally, 10 molar equivalents (relative to maleimide) of L-cysteine ​​(Melford) (dissolved in PBS pH 7.4) were added, and the mixture was incubated for 15 minutes to complete the reaction.

[0220] Alternatively, 1, 2, 4, or 8 molar equivalents (relative to protein thiols) of Alexa488 C5 maleimide (ThermoFisher Scientific, dissolved in DMSO), MC-Val-Cit-PAB-MMAF, Mal-PEG8-Val-Cit-PAB-MMAF, Mal-PEG8-Val-Cit-PAB-MMAE, or MC-PEG8-Val-Ala-PAB-exatecan (MedChemExpress, dissolved in DMSO) were added to the reduced protein, and the mixture was incubated at 25°C for 2 hours. Finally, 10 molar equivalents (relative to maleimide) of L-cysteine ​​(Melford) (dissolved in PBS pH 7.4) were added, and the mixture was incubated for 15 minutes to complete the reaction.

[0221] Purification of dye-protein conjugates The fluorescein-5-maleimide reaction mixture was centrifuged at 17,000 g for 10 minutes to remove precipitate. The reaction mixture was then loaded onto a PD-10 gravity-flow desalting column (Cytiva), followed by loading of up to 2.5 mL of PBS buffer. The conjugate protein was eluted with an additional 3.5 mL of PBS buffer. A 250 μL elution fraction with high protein concentration was selected by UV / Vis spectrophotometry using an IMPLEN nanophotometer at wavelengths of 280 nm and 495 nm, and concentrated to a final volume of approximately 100–400 μL using a 0.5 mL Pierce® protein concentrator (ThermoFisher) with either a 3 or 10 kDa MWCO filter. Protein concentration was adjusted using the following formula for the dye's contribution to absorbance at 280 nm.

[0222]

number

[0223] A 280 This refers to the absorbance of the dye (payload)-protein conjugate at 280 nm, and A 495 This refers to the absorbance at 495 nm, and εprot Refers to the extinction coefficient of the reduced protein at 280 nm calculated using Expasy ProtParam.

[0224] Alternatively, the Alexa488 C5 maleimide reaction mixture was centrifuged at 17,000 g for 10 minutes to remove the precipitate. Subsequently, the reaction mixture was loaded onto a 40K MWCO Zeba™ Spin Desalting Column (ThermoFisher Scientific). Then, the spin desalting column was centrifuged at 700 g according to the manufacturer's instructions to recover the purified sample. The absorbance of the conjugated protein was measured by UV / Vis spectrophotometry using an IMPLEN nanophotometer at wavelengths of 280 nm and 494 nm. The protein concentration was adjusted for the contribution of the dye to the absorbance at 280 nm using the following equation.

[0225]

Equation

[0226] A 280 refers to the absorbance of the dye (payload)-protein conjugate at 280 nm, and A 494 refers to the absorbance at 494 nm, and ε prot refers to the extinction coefficient of the reduced protein at 280 nm calculated using Expasy ProtParam.

[0227] The protein concentration can be estimated to be either monomeric or multimeric depending on the extinction coefficient used for the protein.

[0228] Estimation of dye / protein ratio The final dye / protein ratio of the fluorescein-5-maleimide conjugate after concentration was estimated using the following equation.

[0229]

Equation

[0230] A 495 refers to the absorbance of the dye (payload)-protein conjugate at 495 nm, and εdye refers to the absorption coefficient of the dye at 495 nm (68,000 M -1 cm -1 -1), and C prot refers to the adjusted protein concentration according to the above formula.

[0231] Alternatively, the final dye / protein ratio of the Alexa488 C5 maleimide conjugate was estimated using the following formula.

[0232]

Equation

[0233] A 494 refers to the absorbance of the dye (payload)-protein conjugate at 494 nm, and ε 色素 refers to the absorption coefficient of the dye at 494 nm (71,000 M -1 cm -1 -1), and C prot refers to the adjusted protein concentration according to the above formula.

[0234] The dye / protein ratio can be estimated for either the monomeric or multimeric concentration of the protein.

[0235] Purification of the drug-protein conjugate The reaction mixture was centrifuged at 17,000 g for 10 minutes to remove precipitate. The reaction mixture was then loaded onto a PD-10 gravity-flow desalting column (Cytiva), followed by loading of up to 2.5 mL of PBS buffer. The conjugate protein was eluted with an additional 3.5 mL of PBS buffer. High-protein concentration elution fractions were selected by UV / Vis spectrophotometry using an IMPLEN nanophotometer at 280 nm and wavelengths with high absorbance from the drug. These fractions were then concentrated to a final volume of approximately 100–400 μL using a 0.5 mL Pierce® protein concentrator (ThermoFisher) with a 10 kDa MWCO filter.

[0236] [Table 4]

[0237] Regarding deruxtecan, an approximate deruxtecan analog (compound 21a) 6 ) 280nm (10,764M -1 cm -1 ) and 370nm (20,982M -1 cm -1 The extinction coefficients at ) were used. For DM1, the extinction coefficients at 280 nm and 252 nm determined in (Fishkin, N. Maytansinoid-BODIPY Conjugates: Application to Microscale Determination of Drug Extinction Coefficients and for Quantification of Maytansinoid Analytes. Mol. Pharmaceutics 12, 1745-1751 (2015)) were used. * The absorbance of MMAF at 214 nm was used solely to determine the fraction in which free MC-Val-Cit-PAB-MMAF begins to elute, as the absorbance of MMAF overlaps with that of the protein peptide bond.

[0238] The protein concentration was adjusted using the following formula published in (Chen, Y. Drug-to-Antibody Ratio (DAR) by UV / Vis Spectroscopy. in Antibody-Drug Conjugates (ed. Ducry, L.) 267-273 (Humana Press, 2013)) for the contribution of the drug to the absorbance at 280 nm.

[0239]

Equation

[0240] A 280 refers to the absorbance at 280 nm of the linker-drug-protein conjugate, and A λ薬物 refers to the absorbance at the drug-specific wavelength of the linker-drug-protein conjugate.

[0241]

Equation

[0242]

Equation

[0243]

Equation

[0245] Subsequently, the concentrated sample was further purified by dialysis in 14 ml of PBS buffer for at least 2 hours each, using a Slide-A-Lyzer® minidialysis device (Thermo Scientific, sample volume 0.5 ml, 20 kDa MWCO cutoff filter). The final linker-drug-protein conjugate concentration was determined using the Coomassie-plus-Bradford assay (ThermoFisher).

[0246] Alternatively, the reaction mixture was centrifuged at 17,000 g for 10 minutes to remove precipitate. The reaction mixture was then loaded onto a 40K MWCO Zeba® spin desalting column (ThermoFisher Scientific). The spin desalting column was then centrifuged at 700 g according to the manufacturer's instructions to collect the purified sample.

[0247] Next, the purified sample was transferred to a 3 mL Slide-A-Lyzer® 20K MWCO dialysis cassette or a 0.5 mL Slide-A-Lyzer® 20K MWCO dialysis device in a 15 mL Falcon tube (ThermoFisher Scientific), and dialyzed for at least 3 rounds according to the manufacturer's instructions for use.

[0248] Alternatively, after the initial spin desalting step, the purified sample was loaded onto another 40K MWCO Zeba™ spin desalting column and centrifuged at 700g according to the manufacturer's instructions to recover the purified sample. Protein concentrations were first estimated using the formula above. For conjugates with MC-Val-Cit-PAB-MMAF, Mal-PEG8-Val-Cit-PAB-MMAF, and Mal-PEG8-Val-Cit-PAB-MMAE, the λ at 248 nm was used. 薬物 , 15,900M -1 cm -1 of

[0249]

number

[0250]

number

[0251] Protein concentration can be estimated as either monomers or polymers, depending on the extinction coefficient used for the protein.

[0252] Estimation of payload / protein ratio The final drug concentration after dialysis was estimated using the following formula.

[0253]

number

[0254] Next, the final drug / protein molar ratio is determined.

[0255]

number

[0256] The drug / protein ratio can be estimated based on the concentration of either monomers or polymers of the protein.

[0257] LC-MS The molecular weights of the dye-protein and drug-protein conjugates SpyCatcher3-G4S-HsCutA1[44~179, C75A, C96A, K82C]-G4S-DogCatcher, or CC042 (SEQ ID NO: 38) were analyzed using LC-MS. An Agilent 1290 UPLC and 6550 ESI Q-ToF mass spectrometer were used. LC buffer A was 0.1% formic acid in water, and LC buffer B was 0.1% formic acid in acetonitrile. An Agilent PLRP-S 1000A (5 μm 50 × 2.1 mm) column was used at 60°C with a flow rate of 0.4 ml / min and a sample injection volume of 0.2 μL. The solvent composition was maintained at 20% B for 5 minutes, increased to 75% B over 1 minute, then increased to 100% B over 4 minutes, and maintained for 2 minutes. Cation spectra were obtained at a rate of one spectrum per second over the m / z range of 300 to 3200.

[0258] SEC-HPLC The size and aggregate content of some catcher-core and linker-payload conjugate catcher-core samples were analyzed using SEC-HPLC. An Agilent 1100 HPLC was used with a MAb-Pac® SEC-1 column (ThermoFisher Scientific). The LC buffer was 1×PBS pH 7.4, the flow rate was 0.2 ml / min, and the injection volume was 10–15 μL. Data were collected over a 25-minute period.

[0259] Protein quantification by BCA assay Prior to conjugation, samples were quantified using a BCA protein assay kit (ThermoFisher) according to the manufacturer's instructions. BSA standards were diluted with 1×PBS. Purified proteins were diluted to 1 / 5, 1 / 10, and 1 / 20 with 1×PBS to ensure concentrations were within the linear range of the assay. After incubation with the BCA reagent, absorbance at 562 nm was measured using a BMG FluoSTAR Omega plate reader. Concentrations were interpolated in mg / mL based on a standard curve.

[0260] Catcher base protein assembly SpyTag / DogTag-Ligand Assembly Testing (CC041, CC042, CC060, CC068) Reactions of CutA1-based cores fused with SpyCatcher003 and DogCatcher, as well as SpT3 and DgT conjugate proteins, were prepared using 10 μM catcher-core monomers. The molar ratio of each conjugate to catcher-core was 2:1. The reaction was prepared in PBS buffer and incubated at 25°C for 1 hour in a 600 rpm plate shaker. Subsequently, 6× SDS loading buffer was added, and the sample was heated at 95°C for 5 minutes to terminate the reaction. The sample was then prepared for visualization by SDS-PAGE.

[0261] SpyTag / DogTag Ligand Assembly for Cell Assays Another variation of the conjugation method is as follows: Samples were prepared at 20 μM using monomer concentrations with a platform:ligand:ligand ratio of 1:1.2:1.2 conjugate-overload, with SpT3 and DgT conjugate proteins conjugated to a CutA1 base core fused with SpyCatcher003 and DogCatcher, respectively. The reaction was prepared in PBS buffer supplemented with 1×cOmplete EDTA-free protease inhibitor and 50 μg / mL penicillin-streptomycin. Samples were incubated at 25°C for 1 hour in the case of SpT3 and DgT conjugates.

[0262] Excess conjugates were removed by loading and washing onto a resin column. (GST catchers were loaded onto the resin column, washed, and then the assembly mixture was loaded.) The assembly was visualized by SDS-PAGE to verify conjugate removal. More specifically, excess ligands were captured from the assembly mixture when incubated with the resin using a MagneGST resin column pre-loaded with GST catchers. For pre-loading onto the resin, GST catchers were mixed in PBS to achieve final concentrations of 2.5 mg of GST-SpyCatcher003 and 2.5 mg of GST-DogCatcher. Assuming a resin binding capacity of 5 mg / mL, an appropriate amount of MagneGST resin was added to achieve complete capture of both proteins. The suspension was repeatedly washed with PBS until the A280 of the supernatant approached baseline. The suspension was maintained on ice as a 50% slurry with PBS until used for assembly purification. Appropriate volumes of MagneGST resin-conjugated GST-DogCatcher and GST-SpyCatcher003 suspensions were added to the conjugation reaction to achieve a final concentration of 16 μM for each GST-catcher. The suspensions were then incubated at room temperature for 4 hours with shaking at 600 rpm. The resin was allowed to precipitate, and the Opentron magnetic module was engaged for 5 minutes. The supernatant was then transferred to a new 96-well plate. The purified conjugates were then quantified by Bradford assay, followed by SDS-PAGE and Coomassie staining for analysis.

[0263] Automated production of catcher-core assemblies Large-scale assembly reactions were performed using a Hamilton Microlab Star as described above, except that tagged conjugate proteins were added to the catcher-core in equimolar ratios to a final concentration of 2–4 μM per monomer. Furthermore, GST-catcher-based purification was not performed. Assemblies were prepared in 96-well plates and then transferred to 384-well low-dead-volume acoustic-certified source plates or 384-well polypropylene acoustic-certified source plates for cell viability / cytotoxicity assays.

[0264] cell culture NCI-N87 (CRL-5822) cells and HCT116 (CCL-247) cells were obtained from ATCC and cultured as usual in RPMI and McCoy, supplemented with 10% FCS and 5% penicillin / streptomycin, respectively. Additional cell lines for screening were obtained from ATCC or ECACC and cultured as usual in appropriate media according to the supplier's recommendations.

[0265] Cell imaging Anti-His & Anti-CutA1 Imaging 8-well chamber slides (Lab-Tek) were coated with 200 μL of poly-L-lysine while gently shaking at room temperature for 1 hour. The slides were washed 3 times with 500 μL of filtered MilliQ water and dried at room temperature for 2 hours. NCI-N87 cells were added to 300 μL of RPMI-1640 medium supplemented with 10% fetal bovine serum on the 8-well chamber slides at a rate of 1.5 × 10⁶ cells. 4Cells were seeded in wells and incubated at 37°C and 5% CO2. After 24 hours, cells were treated with conjugates or complete assemblies at a 200 nM monomer concentration and incubated at 37°C and 5% CO2 for 2 hours and 45 minutes. Cells were washed 2× with DPBS and fixed in 4% paraformaldehyde in PBS at room temperature for 15 minutes. Then, cells were washed 2× for 5 minutes with DPBS-T (0.1% Triton X-100) and blocked in DPBS-T and 5% FCS (blocking buffer) at room temperature for 1 hour. For immunofluorescence staining, cells were treated overnight at 4°C with primary antibodies (a. anti-His antibody, proteintech66005, b. anti-CutA1 antibody, abcam ab192236) diluted 1:500 in blocking buffer. Cells were washed with DPBS-T for 3 × 5 minutes at room temperature, stained with secondary antibodies (a. goat anti-mouse IgG H&L DL488, b. goat anti-rabbit IgG H&L DL488, Abbexa), washed with DPBS-T for 3 × 5 minutes at room temperature, and mounted using a fluoroshild mounting medium containing DAPI (Abcam). Cells were imaged using a Nikon Ti2 inverted confocal microscope, and the images were analyzed using Fiji.

[0266] Fluorescein imaging 8-well chamber slides (Lab-Tek) were coated with 200 μL of poly-L-lysine while gently shaking at room temperature for 1 hour. The slides were washed 3 × with 500 μL of filtered MilliQ water and dried at room temperature for 2 hours. NCI-N87 cells were placed in 300 μL of RPMI-1640 medium supplemented with 10% fetal bovine serum on the 8-well chamber slides, 2.5 × 10⁶ cells per well. 4 Cells were seeded in wells and incubated at 37°C and 5% CO2. After 24 hours, the cells were removed from fluorescein-5-maleimide-labeled catcher cores (CC). * , 100 nM, monomer concentration), single assembly (CC) * :L1, 100nM, monomer concentration; L2:CC * , 300 nM, monomer concentration), or complete assembly (L2:CC) *Cells were treated with L1 (100 nM monomer concentration) or complete assemblies and incubated at 37°C in 5% CO2 for 4 hours. Cells were 2× washed with complete medium, followed by 2× washing with DPBS, and fixed in 4% paraformaldehyde in PBS for 15 minutes at room temperature. Then, cells were washed 3×5 minutes with DPBS-T (0.05% Tween 20) and mounted using Fluoroshield mounting medium (Abcam) with DAPI. Cells were imaged with a Nikon Ti2 inverted confocal microscope, and images were analyzed using Fiji.

[0267] Cell viability assay / cytotoxicity assay 1,000 HCT116 cells / well were seeded into 96-well plates and grown for 24 hours in McCoy medium supplemented with 10% FCS. Cells were then treated or sham-treated with protein assemblies of varying concentrations (0.8–50 nM) containing a panel of 4×4 orthogonally tagged ligands (L2, L4, L5, L6; each SpT or DgT) or corresponding drug conjugate assemblies (CC042-DM1 assembly or CC068-deruxtecan assembly). Scaffolds alone (CC042 or CC068) and PBS-treated cells were used as controls. Cells were then grown for 4 days, and the viability fraction was measured using the MTT assay. 20 μL of 3-(4,5-dimethylthiazole-2-yl)-2,5-diphenyltetrazolium bromide (MTT) at 5 mg / mL in PBS was added to each well containing 200 μL of medium. After a 3-hour incubation, the culture medium was aspirated, formazan was dissolved in 100% DMSO, and the absorbance at 570 nm was read. The viable fraction was normalized to a sham-treated control. Alternatively, cells were seeded in complete growth medium in 384-well black clear flat-bottom imaging plates at appropriate cell density according to the manufacturer's instructions. The following day, cells were treated with various antibody concentrations (0.5–50 nM in dimers) or single-concentration catcher-core assemblies, positive control (no payload), or left untreated (negative control). Drug candidate addition was performed using an Echo550 with catcher-core assemblies in low-dead-volume acoustically certified source plates. Cells were then grown for 5 days, stained with Hoechst 33342 for 1 hour, and analyzed using a Celigo Imaging Cytometer. Cell viability was assessed by quantifying the total amount of Hoechst+ cells, and the results were normalized to positive and negative controls.

[0268] A brief explanation of abbreviated sequence lists Sequence ID 1 shows the amino acid sequence of the monomer of the human CutA1 protein. Sequence ID 2 shows the amino acid sequence of "SpyCatcher," which has a His tag and an optional N-terminal protease site and linker. Sequence ID 3 shows the amino acid sequence of "SpyTag". In this specification, this is also referred to as "SpyTag 001". Sequence ID 4 shows the amino acid sequence of the least reactive "SpyTag". Sequence ID 5 shows the amino acid sequence of "SpyCatcher". In this specification, this is also referred to as "SpyCatcher 001". Sequence ID 6 shows the amino acid sequence of the least reactive SpyCatcher "ΔN1ΔC2". Sequence ID 7 shows the amino acid sequence of "SpyTag 002". Sequence ID 8 shows the amino acid sequence of "SpyCatcher 002". Sequence ID 9 shows the amino acid sequence of "SpyTag 003". Sequence ID 10 shows the amino acid sequence of "SpyCatcher 003". Sequence ID 11 shows the amino acid sequence of "SnoopCatcher". Sequence ID 12 shows the amino acid sequence of "SnoopTag". Sequence ID 13 shows the amino acid sequence of "DogCatcher". Sequence ID 14 shows the amino acid sequence of "DogTag". Sequence ID 15 shows the monomer amino acid sequence of the His-tagged construct H6-SpyCatcher003-HsCutA1-DogCatcher, which has a shortened HsCutA1, similar to Sequence ID 16; "CC1" is also shown. Sequence ID 16 shows the amino acid sequence of the human CutA1 monomer, representing an intermediate shortening between Sequence ID 1 and Sequence ID 18, based on the resolving structure of PDB ID: 2zfh. Sequence ID 17 shows the monomer amino acid sequence of the construct SpyCatcher003-HsCutA1-DogCatcher, which is used for structure prediction. Sequence ID 18 shows the amino acid sequence of the human CutA1 monomer, resolved to PDB ID:2zfh, which is a shortened version of Sequence ID 1. Sequence ID 19 has HsCutA1 at both ends. 60~171The amino acid sequence of the shortened human CutA1 monomer is shown. Sequence ID 20 shows the amino acid sequence of a human CutA1 monomer similar to Sequence ID 16, which has the C75A and C96A mutations. Sequence ID 21 shows the amino acid sequence of the GGGGS linker fused to HsCutA1 to directly create a fusion construct without using the SpyCatcher, DogCatcher, or SnoopCatcher modules. Sequence ID 22 shows the amino acid sequence of the (GGGGS)3 linker fused to HsCutA1 to directly create a fusion construct without using the SpyCatcher, DogCatcher, or SnoopCatcher modules. Sequence ID 23 shows the amino acid sequence of a human CutA1 monomer similar to Sequence ID 16, which has the C75V and C96S mutations. Sequence ID 24 shows the amino acid sequence of a human CutA1 monomer similar to Sequence ID 19, which has the C75A and C96A mutations. Sequence ID 25 shows the amino acid sequence of a human CutA1 monomer similar to Sequence ID 19, which has the C75V and C96S mutations. Sequence ID 26 has a shortened HsCutA1 similar to Sequence ID 23 and a linker similar to Sequence ID 22, and is a His-tagged construct H6-SpyCatcher003-(GGGGS)3-[C75V,C96S]HsCutA1 44~179 The amino acid sequence of the monomer -(GGGGS)3-DogCatcher is shown. Sequence ID 27 has a shortened HsCutA1 similar to Sequence ID 19, and His-tagged construct H6-SpyCatcher003-(GGGGS)3-HsCutA1 has a linker similar to Sequence ID 22. 60~171 The amino acid sequence of the monomer -(GGGGS)3-DogCatcher is shown. Sequence ID 28 has a shortened HsCutA1 similar to Sequence ID 24 and a linker similar to Sequence ID 22, resulting in the His-tagged construct H6-SpyCatcher003-(GGGGS)3-[C75A,C96A]HsCutA1 60~171The amino acid sequence of the monomer -(GGGGS)3-DogCatcher is shown. Sequence ID 29 has a shortened HsCutA1 similar to Sequence ID 25 and a linker similar to Sequence ID 22, resulting in the His-tagged construct H6-SpyCatcher003-(GGGGS)3-[C75V,C96S]HsCutA1 60~171 The amino acid sequence of the monomer -(GGGGS)3-DogCatcher is shown. Sequence ID 30 shows the amino acid sequence of a human CutA1 monomer similar to Sequence ID 20, which has the E78C mutation. Sequence ID 31 shows the amino acid sequence of a human CutA1 monomer similar to Sequence ID 20, which has the K82C mutation. Sequence ID 32 shows the amino acid sequence of a human CutA1 monomer similar to Sequence ID 20, which has the Q102C mutation. Sequence ID 33 shows the amino acid sequence of a human CutA1 monomer similar to Sequence ID 20, which has the E114C mutation. Sequence ID 34 shows the amino acid sequence of the human CutA1 monomer, similar to Sequence ID 20, which has the F136C mutation. Sequence ID 35 shows the amino acid sequence of a human CutA1 monomer similar to Sequence ID 20, which has the Q166C mutation. Sequence ID 36 has a shortened HsCutA1 similar to Sequence ID 20 and a linker similar to Sequence ID 21, resulting in the His-tagged construct H6-SpyCatcher003-GGGGS-[C75A,C96A]HsCutA1 44~179 The amino acid sequence of -GGGGS-DogCatcher is shown. "CC041" is also shown. Sequence ID 37 has a shortened HsCutA1 similar to Sequence ID 30 and a linker similar to Sequence ID 21, resulting in the His-tagged construct H6-SpyCatcher003-GGGGS-[C75A, C96A, E78C]HsCutA1 44~179 The amino acid sequence of -GGGGS-DogCatcher is shown. "CC056" is also shown. Sequence ID 38 has a shortened HsCutA1 similar to Sequence ID 31 and a linker similar to Sequence ID 21, resulting in the His-tagged construct H6-SpyCatcher003-GGGGS-[C75A, C96A, K82C]HsCutA1 44~179 The amino acid sequence of -GGGGS-DogCatcher is shown. "CC042" is also shown. Sequence ID 39 has a shortened HsCutA1 similar to Sequence ID 32 and a linker similar to Sequence ID 21, resulting in the His-tagged construct H6-SpyCatcher003-GGGGS-[C75A, C96A, Q102C]HsCutA1 44~179 The amino acid sequence of -GGGGS-DogCatcher is shown. "CC057" is also shown. Sequence ID 40 has a shortened HsCutA1 similar to Sequence ID 33 and a linker similar to Sequence ID 21, and is a His-tagged construct H6-SpyCatcher003-GGGGS-[C75A, C96A, E114C]HsCutA1 44~179 The amino acid sequence of -GGGGS-DogCatcher is shown. "CC038" is also shown. Sequence ID 41 has a shortened HsCutA1 similar to Sequence ID 34 and a linker similar to Sequence ID 21, resulting in the His-tagged construct H6-SpyCatcher003-GGGGS-[C75A, C96A, F136C]HsCutA1 44~179 The amino acid sequence of -GGGGS-DogCatcher is shown. "CC044" is also shown. Sequence ID 42 has a shortened HsCutA1 similar to Sequence ID 35 and a linker similar to Sequence ID 21, resulting in the His-tagged construct H6-SpyCatcher003-GGGGS-[C75A, C96A, Q166C]HsCutA1 44~179 The amino acid sequence of -GGGGS-DogCatcher is shown. "CC058" is also shown. Sequence ID 43 shows the amino acid sequence of the wild-type human IgG1 hinge region. Sequence ID 44 shows the amino acid sequence of the human IgG1 hinge region with the C230S (PDB numbering) mutation. Sequence ID 45 shows the amino acid sequence of the (GGGGS)2 linker fused to HsCutA1 to directly create a fusion construct without using the SpyCatcher, DogCatcher, or SnoopCatcher modules. Sequence ID 46 shows the amino acid sequence of the non-hinge wild-type human IgG1 Fc region. Sequence ID 47 shows the amino acid sequence of the construct SpyCatcher003-hinge[C230S]-Fc-(GGGGS)2-DogCatcher, which has a hinge region similar to Sequence ID 44 and a C-terminal linker similar to Sequence ID 45. Sequence ID 48 shows the amino acid sequence of the construct H6-SpyCatcher003-(GGGGS)3-Fc-(GGGGS)3-DogCatcher, which has a linker similar to that of Sequence ID 22. Sequence ID 49 shows the amino acid sequence of His-GST-(GGGGS)3-SpyCatcher003. Sequence ID 50 shows the amino acid sequence of His-GST-(GGGGS)3-DogCatcher. Sequence ID 51 is the His-tagged construct H6-SpyCatcher003-(GGGGS)3-[C75V,C96S]HsCutA1 60~171 The amino acid sequence of the monomer -(GGGGS)3-DogCatcher is shown. Sequence ID 52 shows the monomer amino acid sequence of the CC076 construct SpyCatcher003-(GGGGS)3-DogCatcher-hinge[C230S]-Fc. Sequence ID 53 shows the monomer amino acid sequence of the CC080 construct SpyCatcher003[S49C]-hinge-Fc-(GGGGS)2-DogCatcher. This construct has a wild-type hinge region sequence similar to Sequence ID 43, and a total of four Cys residues per monomer in the SpyCatcher003 domain with [S49C] mutations. Sequence ID 54 shows the monomer amino acid sequence of the CC086 construct SpyCatcher003[S49C]-GGGGS-hinge-Fc-(GGGGS)2-DogCatcher. This construct has a wild-type hinge region sequence similar to Sequence ID 43, and [S49C] mutations in a total of four Cys residues per monomer in the SpyCatcher003 domain.

[0269] [ka]

[0270] [ka]

[0271] [ka]

[0272] [ka]

[0273] [ka]

[0274] Alignment showing sequence variation:

[0275] [ka]

[0276] [ka]

[0277] References for the Examples [Table 5]

[0278] The examples and embodiments described herein are for illustrative purposes only, and those skilled in the art will understand that various modifications or changes in light of them are suggested and will be included in the spirit and scope of this application and the appended claims. All publications, sequence acceptance numbers, patents, and patent applications referenced herein are incorporated herein by reference in their entirety for all purposes.

Claims

1. A method for preparing a polyvalent bound polypeptide conjugated to a target molecule, The method comprises the step of combining a first polypeptide conjugated to a target molecule with at least one other polypeptide to form a polyvalent polypeptide conjugated to the target molecule.

2. (a) the step of conjugating the target molecule to the first polypeptide of the polyvalent bonded polypeptide; and (b) The step of combining the conjugated first polypeptide of the polyvalent polypeptide obtained in step (a) with at least one other polypeptide of the polyvalent polypeptide to form the polyvalent polypeptide conjugated into the target molecule. A method for preparing a polyvalent polypeptide conjugated to a target molecule, according to claim 1, comprising:

3. The target molecule contains or consists of a drug, label, or dye, and optionally the target molecule is a target nonprotein molecule, optionally a nucleic acid; and / or The polyvalent polypeptide is bispecific or triplicate; and / or The polyvalent conjugated polypeptide is a bispecific antibody; and / or The polyvalent conjugated polypeptide conjugated to the target non-protein molecule is a bispecific antibody-drug conjugate (bsADC). The method according to claim 1 or 2.

4. The aforementioned polyvalent antigen-binding polypeptide is a bispecific antibody, The method according to claim 2 or 3, wherein step (a) is to conjugate the target molecule to a first polypeptide chain of the bispecific antibody, and step (b) is to combine the conjugated first polypeptide chain of the bispecific antibody with at least a second chain of the bispecific antibody to form the bispecific antibody.

5. The first polypeptide of the polyvalent antigen-binding polypeptide comprises or consists of a polypeptide comprising a first binding domain, a second binding domain, and a structural domain. Optional, (a) The first binding domain is located at the N-terminus, the second binding domain is located at the C-terminus, and the first and second binding domains are separated by a structural domain; or (b) The first binding domain is connected to the second binding domain, the second binding domain is connected to the structural domain, and optionally the second binding domain is connected to the N-terminus or C-terminus of the structural domain. The method according to any one of claims 1 to 3.

6. The method according to claim 5, wherein the target molecule is optionally conjugated in the structural domain or in the linker region between the structural domain and the binding domain in step (a) of claim 2.

7. The first binding domain and the second binding domain are catcher domains capable of forming isopeptide links with homologous peptides, Optionally, the homologous peptide for the first binding domain is different from the homologous peptide for the second binding domain, or Optionally, the homologous peptide for the first binding domain is the same as that for the second binding domain, and optionally, the first binding domain and / or the second binding domain are capable of selective binding or conjugation by temporal or sequential control, such as activation or inactivation of at least one binding domain, and / or competitive binding. The method according to claim 5 or 6.

8. The method according to any one of claims 5 to 7, wherein the structural domain is CutA1 or a variant thereof.

9. The method according to claim 7, wherein the structural domain is an Fc region or a variant thereof.

10. The method according to any one of claims 7 to 9, wherein the at least one other polypeptide comprises an antigen-binding domain having a homologous peptide capable of forming an isopeptide bond with at least one of the catcher domains.

11. The method according to any one of claims 7 to 10, wherein the at least one other polypeptide comprises two separate antigen-binding polypeptides, each having a homologous peptide capable of forming an isopeptide bond with one of the catcher domains, and optionally, the two separate antigen-binding polypeptides are combined simultaneously or sequentially with the first polypeptide.

12. The method described above is A further step, optionally obtained from step (b) of claim 2, wherein the polyvalent binding polypeptide is modified to remove the catcher domain, thereby providing an antigen-binding molecule comprising first and second antigen-binding domains separated by a structural domain. The method according to claim 11, including the method described in claim 11.

13. The method according to claim 11, as dependent on claim 2, wherein a plurality of conjugated first polypeptides of the polyvalent antigen-binding polypeptide are generated in step (a), and in step (b), are exposed to a plurality of different pairs of antigen-binding polypeptides, each having a homologous peptide capable of forming an isopeptide bond with one of the catcher domains, resulting in the generation of a variety of different bispecific antigen-binding polypeptides, each conjugated to a non-protein molecule of interest.

14. The method according to claim 2, or any of claims 3 to 13 as dependent on claim 2, wherein the conjugated first polypeptide of the polyvalent polypeptide obtained from step (a) is stored for at least one hour prior to step (b), and optionally the storage period is at least one day, at least one week, at least one month, or at least six months.

15. The storage is performed at a temperature between 10°C and -130°C, optionally between 10°C and -100°C, optionally between 8°C and -80°C, or between 4°C and -20°C, or between 4°C and 0°C. The method according to claim 14.

16. A polyvalent conjugated polypeptide conjugated to a target molecule, which can be obtained or obtained by any method of claim 1 to 15, wherein the polyvalent conjugated polypeptide is bispecific, tripspecific, or has higher-order multiple specificity.

17. A method for preparing a population of multivalent antigen-binding proteins, wherein the members of the population have different antigen-binding domains, and each multivalent antigen-binding protein is conjugated to a non-protein molecule of interest. The aforementioned method, (a) A step of preparing a plurality of polypeptides, each containing a target molecule conjugated thereto, Each polypeptide comprises a first binding domain, a second binding domain, and a structural domain, wherein the first and second binding domains are catcher domains capable of forming isopeptide links with a congener peptide, and the congener peptide for the first binding domain is different from the congener peptide for the second binding domain, and (b) The step of bringing each of the plurality of polypeptides, which include the target small molecule conjugated thereto, into contact with a pair of antigen-binding polypeptides, each having a congener peptide capable of forming an isopeptide bond with one of the catcher domains, under conditions that enable the formation of the isopeptide bond between the catcher domain and the congener peptide, The step involves contacting different pairs of antigen-binding polypeptides with different polypeptides containing the target small molecule conjugated to them, thereby providing a plurality of antigen-binding polypeptides having different antigen-binding characteristics and the same target small molecule conjugated to them. A method that includes this.

18. (c) The method of claim 17, further comprising the step of evaluating one or more features of the plurality of antigen-binding polypeptides having different antigen-binding features and small molecules of the same purpose conjugated thereto.

19. A library of multispecific antigen-binding polypeptides having different antigen-binding features and small molecules of the same purpose conjugated thereto, which can be obtained or may be obtained by the method of claim 17.

20. A polyvalent binding polypeptide conjugated to a target non-protein molecule, wherein the polyvalent binding polypeptide comprises a first binding domain, a second binding domain, and a structural domain, the first binding domain and the second binding domain each being a catcher domain capable of forming an isopeptide linkage with a congener peptide, and the congener peptide for the first binding domain is different from the congener peptide for the second binding domain. Optional, (a) The first binding domain is located at the N-terminus, the second binding domain is located at the C-terminus, and the first and second binding domains are separated by a structural domain; or (b) A polyvalent binding polypeptide in which the first binding domain is connected to the second binding domain, and the second binding domain is connected to the structural domain.

21. The catcher domain is covalently linked to its conjugate peptides by isopeptide linkage, and the conjugate peptides to the first binding domain are different from the conjugate peptides to the second binding domain, wherein each conjugate peptide is covalently attached to a different antigen-binding polypeptide, as described in claim 20.

22. A group of polyvalently bound polypeptides conjugated to the target molecule, as described in claim 20 or 21.

23. A polyvalent polyvalent polypeptide conjugated to a molecule of interest according to claim 20 or 21, or a group according to claim 22, formulated as a composition having one or more excipients of choice for storage at refrigerated or frozen temperatures for at least one hour, at least one day, at least one week, at least one month, or at least six months, wherein the one or more excipients of choice comprise a cryoprotectant.

24. Structure CC068 (Sequence ID 47), or It has the C230S (Fc PDB numbering) mutation, or SpyCatcher-Hinge-Fc-(G4S)2-DogCatcher constructs containing the hinge region and the C230S mutation, or Catcher-Catcher-FC format, or Structure CC076 (Sequence ID 52); or SpyCatcher-Hinge-Fc-(G4S)2-DogCatcher construct (SEQ ID NO: 53) having the wild-type hinge sequence and the S49C mutation in SpyCatcher; or SpyCatcher-GGGGS-Hinge-Fc-(G4S)2-DogCatcher construct (SEQ ID NO: 54) possessing the wild-type hinge sequence and the S49C mutation in SpyCatcher. A structure having at least 85%, at least 90%, at least 95%, at least 96%, at least 97%, at least 98%, or at least 99% identity with sequence numbers 47, 52, 53, or 54. including or consisting of these A polyvalently bound polypeptide conjugated to the target molecule according to claim 20 or 21.

25. A kit comprising, according to claim 20, a plurality of polyvalent conjugated polypeptides conjugated to a target molecule, and a plurality of antigen-binding polypeptides, each having a homologous peptide capable of forming an isopeptide bond with one of the catcher domains.

Images